United States
Environmental
Protection
Agency
Office* Air Quality
Planning and Standards
neasarch Triangle Park, NC 27711
EPA-450/4-90-004
JANUARY 1990
SEPA
AIR
Receptor Model
Technical Series,
Volume III (1989 Revision)
CMB7 User's Manual
-------
EPA-450/4-90-004
RECEPTOR MODEL TECHNICAL SERIES,
VOLUME III C1989 REVISION)
CMB7 USER'S MANUAL
By
John G. Watson
Norman F. Robinson
Judith C. Chow
Desert Research Institute
Reno, Nevada 89506
Ronald C. Henry
Bongmann Kim
University of Southern California
Los Angeles, California
and
Quang T. Nguyen
Edwin L. Meyer
Thompson G. Pace
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Contract No. CX-813187-01-1
EPA Project Officer: Thompson G. Pace
Quang T. Nguyen
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
January 1990
-------
This report has been reviewed by the Office of Air Quality Planning and Standards,
U. S. Environmental Protection Agency and approved for publication as received from
the contractor. Approval does not signify that the contents necessarily reflect the
views and policies of the Agency, neither does mention of trade names or commercial
products constitute endorsement or recommendation for use.
EPA-450/4-90-004
-------
ABSTRACT
The Chemical Mass Balance (CMB) receptor model uses chemical
concentrations measured in source and receptor samples to estimate the
contributions of different source types to ambient pollutant concentrations.
The model is used primarily in the development of State Implementation Plans
for PM10. CMB7 is a software package which implements the model. This
interactive software operates on IBM compatible microcomputers and allows the
user to: 1) select samples, chemical species, and source types for modeling;
2) calculate source contributions and their standard errors using the
effective variance least squares estimation algorithm; 3) evaluate the
goodness of fit and validate the model results; 4) prepare outputs for reports
and input to data base and spreadsheet software; and 5) graph results.
The User's Manual describes model installation, the command menu, and
the structure of input and output data files. A step-by-step tutorial using
example data files, which are included on a floppy disk with the executable
computer code, is provided. Sources of ambient and source profile data for
PMio are described.
-------
ACKNOWLEDGEMENTS
The authors are Indebted to many individuals who have used this software
and contributed suggestions for its improvement. These individuals include,
but are not limited to, Mr. John Core, Mr. Patrick Hanrahan, Dr. fhung Liu,
Dr. Andrew Gray, Dr. John Cooper, Dr. James Huntzicker, Dr. Luke Wijnberg, Dr.
Jitendra Shah, Mr. David Maughan, Ms. Karen Magliano, and Mr. Chuck Unger.
Model development and documentation were partially funded by a
cooperative agreement with the U.S. Environmental Protection Agency, No. CX-
813187-01-1.
-------
CONTENTS
Pace
ABSTRACT i
ACKNOWLEDGEMENTS i i
LIST OF FIGURES iv
LIST OF TABLES iv
1.0 INTRODUCTION 1
1.1 What The Model Does 2
1.2 How The Model Works 4
1.3 CMB Software History 5
1.4 Differences Between CMB7 and CMB 6.0 7
1.5 Organization of User's Manual 8
2.0 SOFTWARE INSTALLATION 11
2.1 Hardware and Operating System 11
2.2 CMB Software 11
2.3 Software Installation 14
3.0 CMB MODEL OPERATIONS , 16
3.1 CMB Commands 16
3.2 Example CMB Application 22
4.0 INPUT AND OUTPUT DATA FILES 41
4.1 CMB7 Input and Output File Descriptions 41
4.2 Creating Data Input Files 49
4.3 CMB 6.0 Files Formats 51
5.0 CMB PERFORMANCE MEASURES 55
5.1 Source Contribution Estimates Display 55
5.2 Similarity/Uncertainty Cluster Display 57
5.3 Species Concentrations Display 58
5.4 Additional Diagnostics 59
6.0 SOURCE AND RECEPTOR PARTICULATE DATA BASES FOR THE CMB.... 61
6.1 Data Base Requirements 61
6.2 Data Base Survey 62
7.0 REFERENCES 76
APPENDIX A - Theory of the Chemical Mass Balance
Receptor Model A-1
APPENDIX B - CMB7 Error Messages and Corrective
Act i on B-1
APPENDIX C - Printout of Test Data Set For PACS1 C-l
iii
-------
-------
LIST OF FIGURES
Page
Figure 1 CMB7 Model Inputs and Outputs 3
Figure 2 Example of Bar Chart from CMB Graphics Menu 36
Figure 3 Example of Pie Chart from CMB Graphics Menu 37
LIST OF TABLES
Page
Table 1 Summary of CMB Source Profiles 64
Table 2 Ambient Particulate Data Bases 67
iv
-------
SECTION 1
INTRODUCTION
The Chemical Mass Balance (CMB) air quality model is one of several
receptor models which have been applied to air resources management. Receptor
models use the chemical and physical characteristics of gases and particles
measured at source and receptor to both identify the presence of and to
quantify source contributions to receptor concentrations. Receptor models are
generally contrasted with dispersion models which use estimates of pollutant
emissions rates, meteorological transport, and chemical transformation
mechanisms to estimate the contribution of each source to receptor
concentrations. The two types of models are complementary, with each type
having strengths which compensate for the weaknesses of the other. The current
guidance for the development of PMj0 State Implementation Plans (SIPs)
recommends the application of both receptor and dispersion models with a
reconciliation of their independent source apportionments (U.S. EPA, 1987c).
This software manual updates Volume III of EPA's Receptor Model
Technical Series by describing the fundamentals and use of CMB7. Volume III
(U.S. EPA, 1987a) describes the earlier CMB 6.0 of the CMB software.
The primary objectives of this manual are:
• To identify the new features and improvements in CMB modeling
software.
• To provide instructions in the use of the new software.
• To identify additional examples of CMB applications to PMjQ,
visibility, and gaseous species.
This manual is intended for wide use by state and local air pollution
control agency personnel in developing State Implementation Plans for PMjQ
(suspended particulate matter with an aerodynamic diameter less than 10 /an).
-------
The U.S. Environmental Protection Agency (EPA) has published two companion
documents to this manual that are also SIP-oriented and that should be
consulted for SIP development. The first, "Protocol for Applying and
Validating the CMB Model" (U.S. EPA, 1987b), provides guidance on
applicability, assumptions, and interpretation of results. The protocol
provides a practical strategy for obtaining valid results. The second
document, "Protocol For Reconciling Differences Among Receptor and Dispersion
Models" (U.S. EPA, 1987c), recommends a procedure for examining and
reconciling differences between receptor and dispersion modeling results.
This manual is not intended to describe fully the CMB or other receptor
models or their applicability to different situations. Several review
articles, books, and conference proceedings provide additional information
about the CMB and other receptor models (Hopke and Dattner, 1982; Stevens and
Pace, 1984; Hopke, 1985; Pace, 1986; Gordon, 1980, 1988; Watson, 1989).
1.1 WHAT THE MODEL DOES
The CMB model uses the chemical composition of ambient pollution samples
to estimate the contributions of different source types to the measured
pollutant concentrations. The CMB model has been most widely used for
suspended particulate matter, but it is equally applicable to gaseous species.
The chemical composition of each source-type's emissions (source profile) must
also be known to use the model. The information required by and produced by
the CMB model is shown in Figure 1.
The CMB model quantifies contributions from chemically distinct source-
types rather than contributions from individual emitters. For example, the
model might calculate that 6.7 ± 2.2 /«}/m3 of PM1Q are contributed by residual
oil combustion, but this contribution might not be further resolved into
concentrations attributable to Power Plant 2, Industrial Boiler 3, Hospital
Heating Plant 6, etc. Sources which have similar chemical compositions cannot
be separated by the model. The software performs tests on ambient data and
source profiles which tell how well source-type contributions can be resolved
from each other.
The CMB7 software can be applied to measurements of up to four different
particle size fractions when source and ambient data are available. When
-------
Receptor
Concentration
File
Source
Profile
File
INPUT FILES
CMB7
Software
OUTPUT
Source
Contribution
Report File
FILES
Source
Contribution
Data Base File
HPGL
Plot
Files
Figure 1. CMB7 Model inputs and outputs.
modeling PMj0, the fine (Pt^ 5, 0 to 2.5 fm aerodynamic diameter) and coarse
(2.5 to 10 /an aerodynamic diameter) size fractions are often modeled
separately. This feature allows particle size as well as chemical
characteristics to be used in distinguishing one source type from another.
The CMB model calculates source contribution estimates for each
individual ambient sample, which is usually of 24-hour duration for PMj0. The
chemical profiles may differ from one sample to the next owing to differences
in emission rates (e.g., some days may have wood-stove burning bans in effect
and others will not), wind directions (e.g., a downwind point source would not
be expected to be contributing at an upwind sampling site), and changes in
emissions compositions (e.g., different gasoline characteristics and engine
performance in winter and summer may result in different profiles). For PM
-------
and every sample is apportioned separately. Seasonal and annual averages of
source contribution estimates are calculated to evaluate the annual PM10
standard.
The CMB model is normally used to apportion particulate matter which is
directly emitted. The remaining sulfate, nitrate, and organic compounds which
are not attributed to primary emissions are secondary species and are not
attributed directly to sources. Research is currently underway to calculate
or measure fractionated source profiles which simulate the profile as it would
look at the receptor after aging, transformation, deposition, and transport.
When this research is completed, the CMB model may be applicable to the
attribution of secondary as well as primary species.
1.2 HOW THE MODEL WORKS
The CMB model is derived from physical principles with
assumptions stated in Appendix A. Therefore, the CMB is not purely a
statistical model, though the least-squares estimation method used to solve
the CMB equations bears resemblance to multiple linear regression analysis.
The CMB consists of the following set of equations:
C, = F.,5, + F)rSr +. .+ F,3Sj . .+ FuSj 1-1..I, J-1..J
where C, = Concentration of species i measured at a receptor site
Fu = Fraction of species i in emissions from source j
Sj = Estimate of the contribution of source j
I = Number of chemical species
0 = Number of source types
These equations have a unique solution only when the number of species
is equal to or greater than the number of sources. Model evaluation studies
show that the greater the number of species, the more precise the
apportionment. These simultaneous equations are solved by an effective
variance weighted least squares estimation method (Watson et al., 1984) which
has been thoroughly tested and documented. This method also requires
-------
variance weighted least squares estimation method (Watson et al., 1984) which
has been thoroughly tested and documented. This method also requires
precision estimates for the C^ and F^j values as model input. These precision
estimates result in realistic uncertainties associated with the source
contribution estimates, S^, which are calculated by the model. All source
J
contribution estimates must be accompanied by their calculated standard errors
when CMB modeling results are reported.
1.3 CMB SOFTWARE HISTORY
The Chemical Mass Balance (CMB) receptor model was first applied by
Winchester and Nifong (1971), Hidy and Friedlander (1972), and Kneip et al.
(1972). The original applications used unique chemical species associated
with each source-type, the so-called "tracer" solution. Friedlander (1973)
introduced the ordinary weighted least-squares solution to the CMB equations,
and this had the advantages of relaxing the constraint of a unique species in
each source-type and of providing estimates of uncertainties associated with
the source contributions.
Gordon and his students at the University of Maryland (e.g., Kowalkzyk
et. al., 1978) subsequently applied this method to many chemical species that
they measured in source and receptor samples. The ordinary weighted least
squares solution was limited in that only the uncertainties of the receptor
concentrations were considered; the uncertainties of the source profiles,
which are typically much higher than the uncertainties of the receptor
concentrations, were neglected.
The first user-oriented software for the CMB model was programmed in
1978 at the Oregon Graduate Center in FORTRAN IV on a PRIME 300 minicomputer
(Watson, 1979). The PRIME 300 was limited to 3 megabytes of storage and 64
kilobytes of random access memory. CMB Versions 1 through 6 updated this
original version and were subject to many of the limitations dictated by the
original computing system. CMB7 has been completely rewritten in a
combination of the C and FORTRAN languages to operate on microcomputers with
floating-point coprocessors, hard disk systems with tens of megabytes storage,
and available memory of 640 kilobytes.
-------
CMB 1 was used in the Portland Aerosol Characterization Study (PACS) to
develop a State Implementation Plan for the control of Total Suspended
Particulate Matter (Watson, 1979). This modeling was the first to identify
and quantify residential wood combustion as a major contributor to particulate
levels in a U.S. urban area.
Version 2 of the CMB software was installed on EPA's UNIVAC system in
1980. This model could be operated by direct dial-up from a remote terminal.
The CMB 2 software was identical to CMB 1 except that the data input files
were generalized. This version was used to introduce state and local
pollution control agencies to receptor modeling in a series of workshops which
were conducted during 1981.
CMB 3 included re-writing of the computer code in FORTRAN 77 and added a
ridge regression solution to the effective variance least-squares estimation
method for solving the CMB equations (Williamson and DuBose, 1983). This
version operated on the EPA UNIVAC via remote terminals. The ridge regression
algorithm was thought to reduce the effects of collinearity (i.e., two or more
source profiles which are too similar to be separated from each other by the
model) on source contribution estimates. Henry (1982) showed, however, that
the ridge regression solution was equivalent to changing the source profiles
from their measured values until the collinearity disappeared. He determined
that the source contribution estimates given by the ridge regression solution
did not represent reality, and its use for air quality modeling was abandoned.
CMB 4, created in 1984, ported the CMB 3 software to an IBM/XT
microcomputer for the first time and added the original effective variance
solution of CMB 1.
CMB Versions 2 through 4 were rarely used for air quality modeling owing
to a lack of appropriate model input data, inadequate user instructions, and
"buggy" software. The anticipation of a revised ambient air quality standard
for suspended particulate matter, PMjg, resulted in a cooperative agreement
between the U.S. Environmental Protection Agency and the Desert Research
Institute of the University of Nevada System to re-package the software for
regulatory applications.
CMB 5 was an experimental version which contained several solution
methods, performance diagnostics, and output displays which could be easily
-------
evaluated by scientists and regulators via application to real and synthesized
data sets. CMB 5 was revised nine times in response to recommendations and
findings of these scientists and regulators. These individuals met in 1986 to
finalize the software for regulatory applications and to prepare instructions
for the use of this software in these applications. CMB 6.0 incorporated the
recommendations of this group into software, a user's manual (U.S. EPA,
1987a), a protocol for applying and validating the CMB model (U.S. EPA,
1987b), and a protocol for reconciling CMB source apportionments with source
apportionments determined by dispersion modeling (U.S. EPA, 1987c).
Even as CMB 6.0 was issued, its limitations were apparent. The computer
code had never been intended for the capabilities of microcomputers. The data
interfaces were inflexible and inconvenient, the memory allocation was
inefficient, and the computational algorithms were slow. Advances in
microcomputer memory, displays, and data analysis software created demands for
more informative displays and easily transferrable model output formats.
These limitations did not prevent the software from being applied in numerous
PMjQ source assessments which are described later in this update. Many
recommendations were made by the users of CMB 6.0, and several software "bugs"
(none of which were found to affect the source contribution estimates) were
found. The user suggestions were compiled and have been addressed in CMB7.
1.4 DIFFERENCES BETWEEN CMB7 AND CMB 6.0
Many of the differences between CMB 6.0 and CMB7 software are internal.
The algorithms and computer code have been rewritten in a combination of the C
and FORTRAN languages to speed data access and calculations. The arrays have
been expanded to accommodate a larger number of source types and receptor
species. All of these changes are transparent to the normal user of the
compiled computer code.
Several changes are external and affect the data input files, the output
files, and the way in which the user interacts with the model:
• CMB7 is menu driven. The CMB 6.0 HELP command has been replaced by
this menu. This feature removes the need to memorize command
mnemonics and to enter them each time a model function is desired.
-------
Each set of source profile and receptor chemical concentrations may
now be supplied to the model as a single, blank delimited ASCII
record (data fields are distinguished by a blank between each field).
These records are easily produced by popular data base and
spreadsheet software. The option remains, however, to use CMB 6.0
source profile and ambient input file formats. (See discussions in
Section 4.)
Ambient concentration and source profile ID codes have been changed
from two to six characters. This allows a much larger number of
source profiles and chemical species to be used in the model. Input
files for CMB 6.0 must be modified by inserting four blank spaces
before the species or source-type code for each record in the input
file. The date field has been expanded to eight characters from the
six-character field length in CMB 6.0. (See Section 4 for details.)
The optional INXXXXXX.IN7 input file which includes the names of
other input data files can contain either five or seven filenames. To
run CMB7 with CMB 6.0 input files, five input data filenames are
listed in the same order as in CMB 6.0 with the *.DAT extension. To
input data from constant record ASCII files (e.g., example files
PRPORT.TXT and ADPORT.TXT), filenames in the sixth and seventh lines
identify receptor and source profile files, respectively, and end
with the *.TXT extender. The first five names may be blank or any
filename without the *.DAT extension. (See Section 4 for details.)
CMB7 uses dynamic memory allocation. CMB7 is no longer limited to 21
fitting species and 16 fitting sources as was CMB 6.0. All remaining
memory up to 640K is available for use by the program. If fewer
species are included, more profiles can be included. The precise
number which can be used depends on the memory configuration of a
given computing system. A number appears on the screen at the start
of each session which indicates the size of the work array available
to CMB7. A formula for the size of the work array required is given
in Appendix A.
CMB7 produces two output files: one similar to that of CMB 6.0 and
another which contains the contribution of each source to each
chemical species in a single, blank-delimited record. These records
can be read directly into commonly used spreadsheet, data base,
graphics, and statistical software.
CMB7 produces pie charts of source contributions and bar charts of
measured and calculated species concentrations. These graphics can
be printed from the screen or directed to a file in HPGL language for
high resolution printing or plotting.
CMB7 produces the Modified Pseudo-inverse Matrix (MPIN) diagnostic to
help identify the degree of influence each chemical species
concentration has on the contribution of the corresponding source.
(See Section 3 for details.)
8
-------
1.5 ORGANIZATION OF USER'S MANUAL
Section 1 has stated the objectives of this User's Manual (this is a
self-sufficient document) and identified the major differences between this
version of CMB modeling software and its predecessor. The second section
describes the software on the distributed EPA diskette, computer hardware
requirements, and how to install CMB7 on IBM-compatible computing systems.
The third section presents a tutorial using example files provided on the EPA
disk. The fourth section describes the input and output files and methods of
building input files. The final section describes the meaning and
interpretation of different model outputs. The appendices describe the CMB
model derivation from basic principles, the solution algorithms, model
assumptions, and the results of tests of deviations from model assumptions.
-------
10
-------
SECTION 2
SOFTWARE INSTALLATION
This section describes the hardware requirements, computer programs, and
installation procedures for CMB7.
2.1 HARDWARE AND OPERATING SYSTEM
The minimum requirements for running CMB7 software are:
• IBM PC compatible desktop, portable, or laptop computer
• Available random access memory in excess of 420K
• Single 5^ dual-sided standard floppy disk drive
• DOS Version 2.0 or higher operating system
The recommended hardware configuration is:
• IBM compatible Intel 80286- or 80386-based microcomputer
• 640K random access memory
• Mathematical co-processor (Intel 8087, 80287, or 80387)
• EGA, VGA, or Hercules compatible video graphics adapter
• HP LaserJet II compatible laser printer
• DOS Version 3.0 or higher operating system
• Hard disk with a minimum of 20M bytes capacity
2.2 CNB SOFTWARE
CMB7 software can be acquired from the National Technical Information
Service or it can be retrieved from the EPA's Support Center For Regulatory
Air Models Bulletin Board System (SCRAM BBS). The software was developed
under a cooperative agreement with the Desert Research Institute of the
University of Nevada, an agency of the State of Nevada, and is non-
proprietary. This software is written in the C and FORTRAN computer languages
and is translated to an executable program using Microsoft (Microsoft Corp.,
11
-------
16011 NE 36th Way, Redmond, WA 98073) compilers. The source code is also
available from the authors on written request accompanied by a formatted,
double density 5V1 floppy disk. The standard disk contains the following
files:
• CMB7.EXE: Executable code for CMB7.
• PROFIN.PRG: A dBaseIII+ or dBase IV utility program which selects
data from EPA's Source Composition Library (U.S. EPA, 1988) for
formatting into CMB7 input data files.
• CMBIN7.PRG: A dBase III+ or dBase IV utility program which converts
dBase files into CMB7 input data file formats for source profile and
ambient data.
• PROFILE.DBF: An example dBase source profile data file in EPA Source
Composition Library format. The entire PROFILE.DBF file and its
documentation (U.S. EPA, 1988) may be obtained from the U.S. EPA
Office of Air Quality Planning and Standards.
• ADPORT.DBF: An example dBase ambient data file. This file may be
used as an example format when requesting data from analytical
laboratories,
• INPORT.IN7: An example input data file containing the filenames of
all other example input files.
• POPORT.IN7: An example input data file containing species codes,
mnemonics, and initial fitting species selections. POPORT.DAT is the
CMB6 copy of this file which is used with the PROFIN program.
• SOPORT.IN7: An example input data file containing source profile
codes, mnemonics, and initial fitting profile selections.
• ADPORT.TXT: An example input ambient data file in CMB7 format. This
file can also be produced from the ADPORT.DBF file using the
CMBIN7.PRG program.
• PRPORT.TXT: An example input source profile data file in CMB7
format. This file can also be produced from a PRPORT.DBF file
(created with the PROFIN.PRG program) using the CMBIN.PRG program.
• ELNAM.STR and PROFILE.STR: These are dBase structure files which are
used by the PROFIN.PRG program.
• MSHERC.COM: This utility program from Microsoft Corp. must be run
prior to running CMB7 to obtain graphics output from a Hercules
compatible video adapter.
12
-------
• README.DOC: This file describes updates on software or the user's
manual which are subsequent to the printing of this manual.
It is good practice to make a working copy of these programs on another
floppy disk and to store the original disk in a safe location.
In addition to this non-proprietary software, several additional
software packages have been found useful as a complement to CMB7:
• EPA Source Composition Library (U.S. EPA, Office of Air Quality
Planning and Standards, Air Quality Management Division, Non-Criteria
Pollutant Program Branch, ND-15, Research Triangle Park, NC 27711).
As noted above, this library and its documentation contain many
source profiles compiled from several studies all over the U.S. The
source profile data base is updated periodically, and new profiles
may be submitted by CMB users to the data base at the stated address.
• VEDIT PLUS (CompuView Products, Inc., 1955 Pauline Blvd., Ann Arbor,
MI 48103). This text editor is able to handle the long lines which
can occur in CMB7 input files.
• dBASE III+ or dBASE IV (Ashton-Tate, 20101 Hamilton Ave., Torrance,
CA 90502). This data base program is useful for assembling CMB7
input data files and for examining the OUXXXXXX.DT2 output files.
• GRAFPLUS (Jewell Technologies, 4740 44th St. SH, Suite 203, Seattle,
WA 98116). This software product can be used to print screen
graphic images on dot matrix printers.
• PrintAPlot (Insight Development Corporation, 1024 Country Club Drive,
Suite 140, Moraga, CA 94556). This software product can be used to
print the HPGL graphic files produced by CMB7 on dot matrix printers
and HP LaserJet Series II laser printers.
• Plotter in a Cartridge (Pacific Data Products, 6464 Nancy Ridge
Drive, San Diego, CA 92121). This hardware product can be used to
print the HPGL graphic files produced by CMB7 on HP LaserJet Series
II laser printers.
• Harvard Presentation Graphics (Software Publishing Corporation, P.O.
Box 7210, Mountain View, CA 94039). Publication-quality pie charts,
stacked bar charts, and time series plots can be produced from CMB7
output data using this plotting package.
The identification of commercially available software products does not
constitute an advertisement or an endorsement of them. Data files can be
constructed and the CMB model can be operated without proprietary software.
13
-------
2.3 SOFTWARE INSTALLATION
It is assumed that the computing system has been properly configured
using the instructions provided by its DOS systems manual. These instructions
will vary from one computer to another. On a system without a hard disk, the
systems disk must be inserted into the floppy disk drive to boot the system.
After booting, the disk prompt (usually A> for a floppy disk system or C> for
a hard disk system) appears on the display. All examples given here will
assume that C> is the default prompt.
The default CONFIG.SYS file must be modified to provide sufficient
buffer and file specifications. A 'FILES=14' and 'BUFFER=20' specification
must appear in the CONFIG.SYS file. The values '14' and '20' are minimum
values, and larger values are also acceptable. This modification may be made
in a text editor or by using the following sequence of commands (computer
prompts are in normal type, user responses are in BOLD FACE, and nonprinting
responses are in parentheses):
C> EDLIN CONFIG.SYS (ENTER)
I (ENTER)
FILES = 14 (ENTER)
BUFFERS = 20 (ENTER)
(CONTROL)(BREAK)
E (ENTER)
The sequence above activates the DOS line editor (EDLIN), names the new
file CONFIG.SYS, enters the Insert mode (I), increases the maximum number of
files available to 14 and the number of buffers to 20, exits the Insert mode,
and exits EDLIN. The system must then be rebooted using the following
sequence:
C> (CONTROL)(ALT)(DEL)
On a hard disk system, the software is copied onto the disk using the
following sequence of commands:
C> NO CHB (ENTER)
C> CD CMB (ENTER)
C>CMB>COPy A:*.* (ENTER)
14
-------
This sequence creates a directory entitled 'CMB', enters that directory,
and copies all files from the program floppy disk in drive A to that
directory. Before using the program, the command C>CD CMB should be entered
to assure that operations are being performed in the proper directory.
With a single floppy disk drive, CMB7 may be executed with the program
disk in Drive A from the A> prompt. After the program displays its first
prompt, the program disk may be removed and another disk with the input files
may be placed in drive A. Depending on the size of the input files, this disk
may have limited space for CMB output files.
CMB input and output files may be read from and written to any disk
drive by specifying the name of that drive in the filename. For example,
specifying B:ADPORT.TXT will cause the software to seek this ambient data
input file on drive B. This is especially useful when a system with two
floppy disk drives is used.
The CMB7 software is now installed. The next section presents a
tutorial using the CMB model with the example data files.
15
-------
SECTION 3
CMB MODEL OPERATIONS
This section describes the CMB model commands and illustrates the use of
these commands in a tutorial example. The tutorial uses the example input
data files which are included with the executable code. This example was
drawn from the Portland Aerosol Characterization Study (Watson, 1979), the
first study in which receptor modeling was used for regulatory purposes. This
example is identical to that used in the tutorial of EPA's Technical Series
Volume III (U.S. EPA, 1987a).
3.1 CMB COMMANDS
CMB7 is menu driven, and all commands are invoked by typing the number
corresponding to the desired action when the prompt is given. The following
menu appears in CMB7 after each operation:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
The number of the desired action is entered after the prompt "Enter
selection, Type the line number to select." These commands are described in
the following sub-sections.
16
-------
3.1.1 Change Fitting Species
Fitting species are those which are used in the calculation of source
contribution estimates. Species which are not included in this calculation
are termed floating species. The comparison of calculated and measured values
for floating species is part of the model validation process (U.S. EPA,
1987b). Fitting species should be selected which are major or unique
components of the source-types influencing the receptor concentrations.
The menu displays row number, species code, species name, and fitting
flags for up to four particle-size ranges. Only the first column is used for
a single particle-size range or non-particulate species. The software
designates the first entry in the list as total mass and sets its fitting flag
to 'T'. This first entry is never used as a fitting species. Species for
which data are lacking have fitting flags set to 'M' and are not available as
fitting species. Species are selected or deselected for the current size
fraction by entering the appropriate row number in response to the prompt. A
'*' designates a fitting species while a blank designates a floating species.
The ENTER key is depressed without the entry of a row number to return to the
command menu.
3.1.2 Change Fitting Sources
Fitting sources are those for which source contribution estimates are
calculated. Several profiles may be available which represent the same
source-type, but only one of these is usually used as a fitting source.
Profiles of similar chemical composition are often found to be collinear when
two or more are selected as fitting sources. Fitting sources are selected
which represent the emissions most likely to influence receptor
concentrations.
The menu displays row number, profile code, profile name, and fitting
flags for up to-four particle-size ranges. Only the first column is used for
a single particle-size range or non-particulate species. Profiles are
selected or deselected for the current size fraction by entering the
appropriate row number in response to the prompt. A '*' designates a fitting
profile and a blank designates a non-fitting profile. The ENTER key is
depressed without the entry of a row number to return to the command menu.
17
-------
3.1.3 Select Samples
A subset of samples may be selected for CMB source apportionment with
this command. This is especially useful when only a single size range is
desired, when the "Perform Autofit" command is used, or when the "Present
Computed Averages" command is used.
The menu displays row number, sample ID, sampling date, sample duration,
start hour, and size range. Samples are selected or deselected for the
current size fraction by entering the appropriate row number in response to
the prompt. A '*' designates a fitting sample, while a blank designates a
sample which will be skipped. The ENTER key is depressed without the entry of
a row number to return to the command menu.
3.1.4 Advance to Next Sample
This command places in current memory the data from the next sample on
the "Select Samples" list. The current fitting sample is deselected when this
command is used.
Note: It is important to use Commands 9 and 12 prior to this command to
save the CMB results.
3.1.5 Calculate Source Contributions
An effective variance weighted least squares calculation is applied to
the fitting species and profiles for the selected sample. Source contribution
estimates, their standard errors, model performance measures, and measured and
calculated species concentrations are displayed to allow an evaluation of the
validity of the solution. Pace and Watson (U.S. EPA, 1987b) explain the use
of the displayed performance measures and methods to validate a source
apportionment calculation. This command may be invoked as many times as
needed to test the effects of selecting different fitting species and
profiles.
Note: Data are recorded on disk only when Commands 9 or 12 are invoked.
3.1.6 Perform AutofIt
Autofit allows a single selection of fitting species and profiles to be
applied to a selected list of samples without operator intervention. This
18
-------
feature is especially useful for model simulation testing and screening
purposes. Autofit is equivalent to invoking Commands 5, 9, 12, and 4 in
succession until the source contributions for all selected samples have been
calculated.
3.1.7 Present Data
This command displays the same information calculated by Command 5
without recalculating the source contribution estimates. The data are
displayed on three separate screens, and this command allows data in these
screens to be reviewed without the delay caused by the least squares
calculation. Command 5 should be invoked at least once before this screen is
displayed for a given selection of fitting species, fitting profiles, and
samples to assure that correct values are displayed.
3.1.8 Present Source Contributions
This command presents a screen display of the fractional contribution of
each source-type to each chemical species. These results are useful when
contributions to species other than total mass are desired. They also
indicate which sources are the major and minor contributors to each chemical
concentration.
3.1.9 Write CMB Information to Disk
This command writes the information displayed by Commands 5 or 7 to the
disk file CMBOUT.DT1. This output can be sent to a printer using the DOS
PRINT command to retain a permanent record of each CMB modeling session. In
addition, a fit to size fraction "COARS" data (Command 5), followed by Command
9, followed by a fit to size fraction "FINE" data without an intervening
change in fitting sources or species, followed again by Command 9, will result
in data being written to CMBOUT.DT1 for size "TOTAL," formed by combining
source contributions for the two size fractions. This is also true for fits
to size fraction "FINE" followed by size fraction "COARS."
Command 9 can be used for every CMB trial to record the changes made
throughout the session. It is usually used only to record the best and final
fit of the sources to the ambient data. A hardcopy record of a working
19
-------
session can be made by simultaneously entering (CTRL)(Print Screen) at the
beginning of the modeling session. The printer must be connected to the
computer and on line. Pressing (CTRL)(Print Screen) a second time switches
off continuous printing. A single screen of information can be printed by
entering (SHIR)(Print Screen).
Note: Not all computers are configured for screen dumps. Check your
computer manual concerning this feature.
3.1.10 Present Computed Averages
This command is ordinarily invoked after Autofit (Command 6) to
calculate averages and standard deviations of the source contributions from a
series of samples. The screen display can be optionally directed to the
OUXXXXXX.DT1 hardcopy file. This command can also be used after a series of
interactive source apportionment calculations. All source contribution
estimates written to disk will be included in the average.
3.1.11 Present Source Profile or Receptor Concentrations
This command produces a screen display of the source profiles and
receptor concentrations. It is used to verify that the input data have been
accurately read into the CMB7 software.
3.1.12 Mrite Source Contributions to Species to Disk
This command produces an output file, entitled OUXXXXXX.DT2, of source
contributions which is convenient for computer manipulation. The output is
not convenient for visual examination. The contribution of each source-type
to each chemical species is written to a single data record for that species.
Each field in the record is blank-delimited, and the file can be easily read
into popular data base or spreadsheet software.
3.1.13 Graph
This command presents a menu for four graph types: 1) bar chart of
calculated and measured receptor species concentrations; 2) bar chart of
source profiles; 3) pie chart of source contributions to a single sample; and
4) pie chart of source contributions to PM,0 (if fine and coarse size
20
-------
fractions have been apportioned). After selection the graph is displayed
until the ENTER key is depressed. Graphs 1 and 2 display species identifiers
along the x axis. When the species are too numerous to be displayed on a
single screen, entering an 'R' will shift the display window to the right and
entering an 'L' will shift the display window to the left. With an appropriate
printer and computer interface, the screen image may be turned into hardcopy
by depressing (SHIFT)(Print Screen) keys. After exiting the graphical display,
the program asks whether a graphics file is desired. A 'Y' response results
in a Hewlett-Packard Graphics Language (HPGL) file being written to disk.
These files can be read directly into desktop publishing and graphics
software.
3.1.14 Present Normalized (over species) MPIN Matrix
This command presents a screen display of the transpose of the
normalized modified pseudo-inverse matrix (MPIN) as described by Kim and Henry
(1989). The MPIN matrix consists of rows for each fitting species and columns
for each fitting source. The entries in this matrix indicate the degree of
influence each species concentration has on the contribution of the
corresponding source and the standard error of that estimate. The normalized
MPIN matrix values indicate the sensitivity of the source contribution and
standard error estimates to individual species. When entries exceed 0.5, the
source contribution is very sensitive to the corresponding species
concentration. The MPIN values are used to determine model stability. If
there is doubt about the accuracy or precision of influential species
concentrations, then the standard errors calculated by the model may not
represent the true uncertainty of the source contributions.
3.1.15 Exit
This command closes all files, terminates program execution and returns
control to the DOS operating system. The program may also be terminated with
a CTRL-C or CTRL-BREAK key sequence, but output data may be lost. The exit
command should be used in all cases except those in which the program appears
to be hung up.
21
-------
3.2 EXAMPLE CMB APPLICATION
This subsection illustrates the use of CMB7 by using the example files
which are supplied with the software. These instructions should be followed
closely during practice sessions with the test data. A printout of the
initial run with the test data set for PACS1 is presented in Appendix C. If
any difficulties are encountered following the instructions in this
subsection, Appendix C should be consulted. Several additional runs for the
same site and sampling date should be performed to determine whether
predictions can be improved by adding or deleting some fitting source(s) or
species.
Instructions in this subsection are in large type while all computer
inputs and outputs are in small type. User responses are in bold-face type.
First, enter the CMB sub-directory
OCO CMB
Start the model by typing
OCMBK3W7
The first software response reminds you that all user input must be in
upper case.
MAKE SURE THAT YOUR CAPS LOCK IS ON ! ! !
You are first asked for the name of the file which contains the names of
the input data files. The contents and formats of these files are discussed
in Section 4.
DISK FILE FOR INITIAL INPUT?
IF NOT ENTER CARRIAGE RETURN
IF SO ENTER NAME OF DISK FILE
IMPORT.IN7
On this prompt, enter the name of the file to which you want your output
data directed. The default name will be CMBOUT. It is good practice to name
these files in such a way that they identify sampling site, season, and
monitoring network. CMB7 supplies its own extensions to this filename, and
any extensions added to this filename are ignored.
DO YOU WISH TO RENAME CMBOUT?
IF NOT ENTER A CARRIAGE RETURN.
IF SO ENTER THE FILE NAME.
OUPORT
22
-------
The model name, date and software authors then appear. Check the EPA's
Support Center For Regulatory Air Models Bulletin Board System (SCRAM BBS) if
you have doubts about whether or not you are operating with the most recent
version.
**************»»*1H»*********«************1«*******^
* *
U. S. EPA CHEMICAL MASS BALANCE RECEPTOR MODEL
• *** IBM-PC CMB7 89338 ***
EPA PROJECT MGRS: THOMPSON G. PACE III, PE
QUANG NGUYEN
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
RESEARCH TRIANGLE PARK, NC
(919)-541-5585
PRINCIPAL AUTHOR: DR. JOHN G. WATSON
DESERT RESEARCH INSTITUTE
UNIVERSITY OF NEVADA SYSTEM
(702)-677-3166
CONTRIBUTING AUTHORS:
DR. J.C. CHOW MR. P.L. HANRAHAN DR. N.F. ROBINSON
MR. J.E. CORE DR. R.C. HENRY DR. H.J. WILLIAMSON
MR. D.A. DUBOSE MR. T.G. PACE DR. L. WIJNBERG
MR. QUANG NGUYEN
**********************
**************************************
71168 (This number specifies the number of separate floating point values
which can be accommodated by the remaining memory in your system. Appendix A
provides a method of relating this value to the number of profiles and species
which you can use in the program.)
The first prompt asks you to
Initialize size fraction by selecting receptor site
Strike enter to continue (ENTER)
PACS1
PACS1
PACS2
PACS2
PACS3
6 PACS3
08/13/77 24 0
08/13/77 24 0
01/24/78 24 0
01/24/78 24 0
08/07/77 24 0
08/07/77 24 0
COARS
FINE
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 1
Enter the line number for each sample. In this case, select the sample
identified on line 1 by typing 1, and the following display appears with an
asterisk opposite line 1. The asterisk indicates that sample is selected for
CMB modeling.
1 PACSl
2 PACSl
3 PACS2
4 PACS2
5 PACS3
6 PACS3
OB/13/77 24 0
08/13/77 24 0
01/24/78 24 0
01/24/78 24 0
08/07/77 24 0
08/07/77 24 0
COARS
FINE
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 2
23
-------
sample.
Select the fine particle sample which corresponds to the coarse particle
1 PACSl
PACS1
PACS2
PACS2
PACS3
6 PACS3
08/13/77 24 0
08/13/77 24 0
01/24/78 24 0
01/24/78 24 0
08/07/77 24 0
08/07/77 24 0
COARS
FINE
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: (ENTER)
Press the ENTER key to return to the main menu.
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 5
Each of the commands described in Section 3.1 is executed by typing its
corresponding line number. By typing '5', for example, the CMB calculation is
made for the first selected sample. There is a short delay (which can
sometimes be up to a minute if operating on an 8086 system without a math
coprocessor or if there are many species and source types), and the source
contribution display appears.
SOURCE CONTRIBUTION ESTIMATES - SITE: PACSl
SAMPLE DURATION 24 START HOUR
R SQUARE .97 PERCENT MASS
CHI SQUARE 1.44 DF
DATE: 08/13/77 CMB7 89338
0 SIZE: COARS
100.6
13
SOURCE
* TYPE
1
3
4
5
8
11
12
13
MARIN
UDUST
AUTPB
RDOIL
KRAFT
ALPRO
STEEL
FERMN
SCE(U6/M3)
10.6029
9.5985
9.0906
9.7127
12.3265
11.0997
8.1587
9.8720
STD ERR
1.8240
1.2616
1.3961
1.6108
7.9102
2.2441
1.5239
1.6165
TSTAT
5.8131
7.6081
6.5112
6.0296
1.5583
4.9462
5.3538
6.1072
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/
Strike enter to continue
160.0+-
11.3
24
-------
This display provides most of the information needed to determine the
source contribution estimates, their standard errors, and the model
performance measures. The interpretation of these measures for each display
is discussed in Section 5 of this manual. To obtain the next display, press
the ENTER key
(ENTER)
and the similarity/uncertainty cluster display appears. This display
indicates those profiles which are similar to each other and might not be
reliably resolved by the CMB model.
UNCERTAINTY/SIMILARITY CLUSTERS CMB7 89338
SUM OF CLUSTER SOURCES
Strike enter to continue
Press ENTER again
(ENTER)
and the species concentration display appears. This display compares the
calculated to measured concentrations of each chemical species. The asterisks
next to a species ID indicate that the species was used in the CMB
calculation.
SPECIES CONCENTRATIONS - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .97 PERCENT MASS
CHI SQUARE 1.44 OF
DATE: 08/13/77 CMB7
0 SIZE:
100.6
13
89338
COARS
SPECIES I — MEAS CALC RATIO C/M RATIO R/U
Cl
C9 -
Cll
C12
C13
C14
C16
C17
CI9
C20
C22
C23
C24
C25
C26
C28
C29
C30
TOT T 80.00000+-
F .73400+-
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
6.33000+-
1.48000+-
4.84000+-
3.27000+-
2.50000+-
4.67000+-
1 . 12000+-
1.52000+-
. 14000+-
.27700+-
.00800+-
2.47000+-
5.41000+-
.77900+-
.05000+-
ZN .21400+-
8.00000
.07300
.63300
. 14800
.48400
.32700
.25000
.46700
.11200
.15200
.01400
.02800
.00100
.24700
.54100
.07800
.00500
.02100
80.46161+-
.50073+-
6.07759+-
1. 48676+-
4.40992+-
3.38783+-
2.41256+-
5.19693+-
1.44197+-
1.47195+-
.13692+-
.34841+-
.24531+-
2.43813+-
4.11644+-
.63016+-
.06613+-
.23739+-
6.36458
.22362
.47089
.52389
. 55081
. 15779
.28216
1 . 07658
.38620
. 10333
.03204
.07289
.11640
.12229
.30571
.11802
. 00683
. 03413
1.01+-
.68+-
.96+-
1.00+-
.91+-
1.04+-
.97+-
1.11+-
1.29+-
.97+-
.98+-
1.26+-
30. 66+- 15
.99+-
.76+-
.81+-
1.32+-
1.11+-
.13
.31
.12
.37
.15
.11
.15
.26
.37
.12
.25
.29
.05
.11
.09
.17
.19
.19
.0
-1.0
-.3
.0
-.6
.3
-.2
.4
.8
-.3
-.1
.9
2.0
-.1
-2.1
-1.1
1.9
.6
Strike enter to continue (ENTER)
Most of the CMB displays provide only a single screen of information at
a time; the remaining information is obtained by depressing ENTER when this
prompt is seen.
25
-------
CSS
C82
C201
C202
C203
C204
BR
PB
OC
EC
S04
N03
1.
10.
1.
8.
1.
52000+-
78000+-
10000+-
68000+-
10000+-
13000*-
.05200
.17800
1.01000
.16800
.81300
.11300
1.
8.
1.
8.
51378+-
93077+-
38185+-
34287+-
11920+-
71048+-
.15795
.27347
1.29686
.39741
1.25659
.41362
1
1
.99+-
.08+-
.83+-
.80+-
.00+-
.63+-
.32
.19
.15
.25
.18
.37
-.0
.5
-1.0
-.8
.0
-1.0
Strike enter to continue
(ENTER)
The main menu appears after each command has been fully executed. This
menu will be abbreviated in the remainder of this tutorial to save space.
1 Change Fitting Species
15 Exit
Type the line number to select: 1
The CMB is an interactive model in which the user must judge which
profiles or species are most appropriate for a given sample. Species may be
eliminated if the measurements are suspect or if an unidentified source is
suspected as being a major contributor to that species. The fitting species
are changed by invoking Command 1.
FINE COARS
T
1 Cl
2 C9
3 Cll
4 C12
5 CIS
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 CSS
20 C82
Sizes:
TOT
F
NA
MG
AL
SI
s
CL
K
CA
TI
V
CR
HN
FE
NI
CU
ZN
BR
PB
F
T
*
*
*
*
*
*
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: U
An asterisk appears next to each species which is used for each fitting
species. The selection shown here is a default which is defined in the
POPORT.IN7 file. The fitting specie's and sources can be different for
different size fractions. There is not enough room on the display for all of
the available species, and entering 'U' shows the remaining portions of the
species 1ist.
26
-------
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 C35
20 C82
21 C201
22 C202
23 C203
24 C204
Sizes
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
N03
FINE COARS
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
Entering an '8' causes the asterisk corresponding to coarse particle CL
to be removed. Chlorine is no longer a fitting species for the coarse
fraction data. The species must appear on the screen to be selected or
deselected.
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 CSS
20 C82
21 C201
22 C202
23 C203
24 C204
Sizes
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
N03
FINE
COARS
*
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: (ENTER)
1 Change Fitting Species
15 Exit'
Type the line number to select: 2 •
27
-------
In the same way, Command 2 allows the fitting source profiles to be
changed interactively.
Sizes: FINE COARS
1 1 HARIN * *
2 2 CDUST
3 3 UDUST * *
4 4 AUTPB * *
5 5 RDOIL * *
6 6 V8RN1
7 7 VBRN2
8 8 KRAFT * *
9 9 SULFT
10 10 H06FU
11 11 ALPRO * *
12 12 STEEL * *
13 13 FERMN * *
14 14 CARBO
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
Entering an '8' causes the asterisk corresponding to the coarse KRAFT
profile to be removed. KRAFT is no longer a fitting profile for the CMB
calculation.
Sizes: FINE COARS
1 1 MARIN * *
2 2 CDUST
3 3 UDUST * *
4 4 AUTPB * *
5 5 RDOIL * *
6 6 VBRN1
7 7 VBRN2
8 8 KRAFT *
9 9 SULFT
10 10 HOGFU
11 11 ALPRO * *
12 12 STEEL * *
13 13 FERMN * *
14 14 CARBO
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: (ENTER)
1 Change Fitting Species
15 Exit'
Type the line number to select: 5
28
-------
Command 5 is invoked to re-calculate the source contribution estimates
after each change in fitting species, profiles, or samples.
«
NO CONVERGENCE AFTER 20 ITERATIONS. ENTER A CARRIAGE RETURN TO VIEW RESULTS
(ENTER)
In this case, the combination of fitting species and sources yields a
very unstable solution. Appendix A shows that the effective variance least
squares estimation method is an iterative solution which converges on the most
probable values. This convergence usually occurs in two to five iterations.
If a solution hasn't converged after 20 iterations, then the selection of
sources and species is not physically meaningful, and other selections must be
made. This message is one of the many internal model validation tests which
are built into CMB7 software. Notice that many of the other performance
measures are outside of reasonable ranges cited in Section 5 in the displays
below.
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .90 PERCENT MASS
CHI SQUARE 4.85 OF
DATE: 08/13/77
0 SIZE:
94.8
12
CMB7 89338
COARS
SOURCE
* TYPE
1
3
4
5
11
12
13
MAR IN
UDUST
AUTPB
RDOIL
ALPRO
STEEL
FERMN
SCE(UG/M3)
12.9724
11.8115
11.1218
11.7584
13.6058 -
-.3121
14.8819
STD ERR
2.0982
1.2227
1.5084
2.0859
2.4641
.2882
1.5638
TSTAT
6.1826
9.6603
7.3734
5.6371
5.5217
-1.0829
9.5168
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/
Strike enter to continue (ENTER)
160.0+- 11.3
UNCERTAINTY/SIMILARITY CLUSTERS CMB7 89338
SUM OF CLUSTER SOURCES
Strike enter to continue (ENTER)
SPECIES CONCENTRATIONS - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .90 PERCENT MASS
CHI SQUARE 4.85 OF
DATE: 08/13/77 CMB7
0 SIZE:
94.8
12
89338
COARS
SPECIES 1 — MEAS CALC —RATIO C/M RATIO R/U
Cl
C9
Cll
C12
C13
C14
C16
C17
C19
C20
C22
TOT T 80.
F
NA
MG
AL
SI
S
CL
K
CA
6.
1.
4.
3.
2.
4.
1.
1.
TI
00000+-
73400+-
33000+-
48000+-
84000+-
27000+-
50000+-
67000+-
12000+-
52000+-
14000+-
8.00000
.07300
.63300
.14800
.48400
.32700
.25000
.46700
.11200
.15200
.01400
75
6
1
5
3
2
5
1
1
.83980+-
.62178+-
.59115+-
. 15164+-
.31604+-
.65624+-
.28332+-
.74261+-
.90394+-
.14509+-
. 14879+-
3.68599
.25880
.56393
.70809
.67496
.19443
.33567
1.30243
.58130
.11261
.02506
.95+-
.85+-
1.04+-
.78+-
1.10+-
1.12+-
.91+-
1.23+-
1.70+-
.75+-
1.06+-
.11
.36
.14
.48
.18
.13
.16
.30
.55
.11
.21
-
-
-
1
-
1
-2
.5
.4
.3
.5
.6
.0
.5
.8
.3
.0
.3
29
-------
C23
C24
C25
C26
C28
C29
C30
C35
C82
C201
C202
C203
C204
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
N03
*•
*
*
*
*
*
*
*
*
*
*
*
.27700+-
. 00800+-
2.47000+-
5.41000+-
.77900+-
.05000+-
.21400+-
.52000+-
1.78000*-
10.10000+-
1 . 68000+-
8.10000+-
1.13000+-
. 02800
.00100
.24700
.54100
.07800
.00500
.02100
.05200
.17800
1.01000
. 16800
.81300
.11300
.41650+-
.01272+-
2.56464+-
1.52394+-
.65911+-
.04401+-
.18275+-
.62465+-
2.28533+-
8.11789+-
1.41114+-
7.95489+-
1.02827+-
. 08825
.00580
.14741
.24114
. 14252
.00719
.04540
.19393
.33450
1.36887
.40161
1.50170
.50149
1.50+-
1.59+-
1.04+-
.28+-
.85+-
.88+-
.85+-
1.20+-
1.28+-
.80+-
.84+-
.98+-
.91+-
.35
.75
.12
.05
.20
.17
.23
.39
.23
.16
.25
.21
.45
1.5
.8
.3
-6.6
-.7
-.7
-.6
.5
1.3
-1.2
-.6
-.1
-.2
Strike enter to continue (EHTER)
1 Change Fitting Species
15 Exit
Type the line number to select:
Re-select CL as a fitting species and KRAFT as a fitting profile by
using Commands 1 and 2 as described above. The screen displays are not
reproduced below.
1 Change Fitting Species
15 Exit'
Type the line number to select: 9
To save results on disk in a hardcopy format which is identical to the
three screen displays, enter Command 9 at the prompt. The screen display is
written to the OUPORT.DT1 data file. Command 12 writes the contribution of
each source profile to each chemical species concentration on a single, long
record into the file OUPORT.DT2. It is a good idea to use both of these
before proceeding to the next sample.
1 Change Fitting Species
15 Exit'
Type the line number to select: 4
Since the coarse fraction apportionment is completed, it is time to
perform the fine fraction apportionment. Command 4 brings the next sample's
data into the model calculation arrays.
1 Change Fitting Species
15 Exit'
Type the line number to select: 5
30
-------
Command 5 once again calculates the source contribution estimates and
presents the three displays of model results.
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .98 PERCENT MASS
CHI SQUARE 1.12 OF
DATE: 08/13/77
0 SIZE:
98.7
13
CMB7 89338
FINE
SOURCE
* TYPE
1
3
4
5
8
11
12
13
MAR IN
UDUST
AUTPB
RDOIL
KRAFT
ALPRO
STEEL
FERMN
SCE(U6/M3)
12.3889
9.5917
10.0835
11.0603
4.6896
10.6023
8.6729
11.8754
STD ERR
2.2457
1.3876
1 . 4942
1.9239
5.0467
3.5896
1.3771
1.8321
TSTAT
5.5167
6.9127
6.7486
5.7490
.9292
2.9536
6.2979
6.4820
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+* 8.0/
Strike enter to continue (ENTER)
160.0+- 11.3
UNCERTAINTY/SIMILARITY CLUSTERS CMB7 89338
SUM OF CLUSTER SOURCES
Strike
1 8
1 5
8
17.078+- 4
28.139+- 3
.241
.833
enter to continue (ENTER)
SPECIES CONCENTRATIONS - SITE: PACS1 DATE: 08/13/77
SAMPLE
DURATION
R SQUARE
CHI SQUARE
24
.98
1.12
START
PERCENT
HOUR 0
MASS . 98.7
OF 13
CMB7
SIZE:
89338
FINE
SPECIES 1 — MEAS CALC RATIO C/M RATIO
Cl
C9
Cll
C12
CIS
C14
C16
C17
C19
C20
C22
C23
C24
C25
C26
C28
C29
C30
C35
C82
C201
C202
C203
C204
TOT T 80
F
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
6
4
.00000+-
.88300+-
.93000+-
.43000+-
.66000+-
3.02000+-
2
5
1
1
2
4
2
7
1
10
N03
.95000+-
.95000+-
.64000+-
.78000+-
. 08300+-
.37200+-
.31500+-
.99000+-
.53000+-
.76500+-
.04400+-
.22500+-
.41900+-
.53000+-
.54000+-
.42000+-
.30000+-
.63800+-
8.00000
.08800
. 69300
. 04300
.46600
.30200
.29500
.59500
. 16400
. 17800
. 00800
.03700
.03200
.29900
.45300
.07700
.00400
.02300
.04200
.25300
.75400
.14200
1 . 03400
.06400
78.96461+- 4
.67644+-
6.97025+-
1.60951+-
4.02418+-
2.92212+-
3.02466+-
5.69381+- 1
1 . 73084+-
1. 43537+-
.10088+-
.39757+-
.20976+-
2.82844+-
4.24446+-
.68246+-
.05274+-
.26786+-
.56133+-
2.13749+-
8.50978+- 1
1. 33579+-
9.78819+- 1
.88402+-
.82449
.24792
.56446
.62627
.88919
.13329
.31807
.24836
.46411
.11366
.01630
.08308
.12151
.14115
.33269
.13428
.00510
.03966
.17386
.30300
.35632
.34012
.47514
.35938
1
3
1
1
1
1
1
1
1
1
1
.99+-
.77+-
.01+-
.74+-
.86+-
.97+-
.03+-
.96+-
.06+-
.81+-
.22+-
.07+-
.67+-
.95+-
.94+-
.89+-
.20+-
.19+-
.34+-
.84+-
.13+-
.94+-
.95+-
.39+-
.12
.29
.13
1.50
.21
.11
.15
.23
.30
.10
.23
.25
.39
.11
.12
.20
.16
.21
.44
.15
.21
.26
.17
.58
R/U
-.1
-.8
.0
1.9
-.6
-.3
.2
-.2
.2
-1.6
1.0
.3
-.8
-.5
-.5
-.5
1,3
.9
.8
-1.0
.6
-.2
-.3
.7
Strike enter to continue (ENTER)
31
-------
In this case, several potential col linearities have been identified by
the similarity/uncertainty clusters. It may be necessary to group several of
these source profiles into a common profile which represents several source
sub-types. No further experimentation will be conducted in this tutorial.
The CMB data are recorded to disk by using Commands 9 and 12 at the menu
prompt.
1 Change Fitting Species
IS Exit'
Type the line number to select: 9 then 12
There are several commands which allow the user to learn more about his
input data and the nature of the CMB least squares calculation. As an
example, invoke Command 3 and select line 1, the PACS1 coarse particle
fraction. Exit to the menu and use Command 5 to re-calculate the source
contributions for these data. The display was presented previously and will
not be repeated.
1 Change Fitting Species
15 Exit
Type the line number to select: 11
Command 11 allows inspection of the data which have been read into the
CMB program. This is useful for verifying that all input data files were
correctly formatted and that the data were correctly retrieved by the
software. It is also useful when a deficit or surplus of a certain species is
found in the species concentrations display. The source profiles can be
examined one by one, or as a group, in order to find profiles which contain
significant quantities of these species.
WHAT DO YOU WANT TO SEE?
ENTER S FOR SOURCE PROFILE OR R FOR RECEPTOR CONCENTRATIONS.
S
DO YOU WANT TO LOOK AT THE WHOLE MATRIX?
IT IS 20 SOURCES BY 24 SPECIES
N
WHICH SOURCE DO YOU WANT?
GIVE SOURCE CODE
4
SOURCE: AUTPB
Cl TOT 1.0000 +- .0000
C9 F .0000 +- .0001
Cll NA .0000 +- .0005
C12 MG .0000 f- .0050
C13 AL .0110 +- .0050
C14 SI .0082 +- .0030
C16 S .0040 -t-- .0013
C17 CL .0300 +- .0100
C19 K .0007 -I- .0003
C20 CA .0125 +- .0050
C22 TI .0000 +- .0010
C23 V .0000 +- .0000
32
-------
C24 CR .0000 +- .0001
C25 MN .0000 +- .0002
C26 FE .0210 +- .0080
C28 NI .0002 +- .0001
C29 CU .0007 +- .0003
C30 ZN .0035 +- .0013
C35 BR .0500 +- .0170
C82 PB .2000 -i- .0300
C201 OC .5000 +- .1000
C202 EC .0380 +- .0140
C203 S04 .0130 +- .0040
C204 N03 .0091 +- .0030
Strike enter to continue
(ENTER)
WHICH SOURCE DO YOU WANT?
GIVE SOURCE CODE
(ENTER)
WHAT DO YOU WANT TO SEE?
ENTER S FOR SOURCE PROFILE OR R FOR RECEPTOR CONCENTRATIONS.
OR ARE YOU DONE?
D
At each of the prompts, enter 'S' for source profiles, 'R' for receptor
data, 'Y' for yes, 'N' for no, or 'D' for done. The 'D' returns to the main
menu.
1 Change Fitting Species
15 Exit
Type the line number to select: 8
Command 8 presents the fractional amount which each source contributes
to each species. This information is also useful when trying to determine
which profiles are responsible for major over- or under-predictions of the
measured species.
CALC SPECIES(PER SOURCE)
INDIVIDUAL RATIO =
MEAS SPECIES(ALL SOURCES)
SPECIES
TOT
F
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
MAR IN
.155
.000
.715
1.383
.000
.000
.139
.833
.106
.097
.000
.000
.000
.000
.000
.000
.000
.000
.059
.000
.000
UDUST
.120
.000
.017
.290
.182
.708
.012
.000
.060
.131
.740
.006
.014
.004
.127
.001
.065
.047
.005
.014
.150
AUTPB
.126
.000
.000
.000
.024
.027
.014
.051
.004
.071
.000
.000
.000
.000
.047
.002
.167
.157
1.203
.797
.669
SOURCE
RDOIL
.138
.007
.056
.000
.013
.035
.499
.000
.019
.098
.147
1.023
.017
.002
.073
.775
.189
.197
.003
.005
.103
CODE
KRAFT
.059
.000
.086
.069
.003
.002
.186
.014
.043
.000
.003
.000
.042
.000
.012
.008
.022
.014
.015
.000
.011
ALPRO
.133
.720
.063
.690
.614
.012
.050
.024
.014
.020
.051
.018
.000
.000
.011
.026
.106
.007
.009
.001
.055
STEEL
.108
.000
.016
1.311
.012
.144
.058
.027
.049
.302
.209
.014
.578
.252
.613
.079
.552
.463
.000
.026
.000
FERMN
.148
.039
.053
.000
.016
.039
.068
.008
.760
.087
.066
.008
.016
.687
.055
.000
.097
.306
.045
.002
.142
33
-------
EC .000 .125 .270 .241 .007 .172 .000 .125
S04 .120 .004 .013 .517 .182 ,045 .021 .048
N03 .000 .000 .144 .113 .000 .068 .000 1.061 '
Strike enter to continue (ENTER)
1 Change Fitting Species
15 Exit
Type the line number to select: 14
Command 14 presents the transpose of the normalized MPIN matrix. This
matrix shows the relative influence of each species on each source
contribution. The species with the largest amount of influence is assigned a
value of ±1, and species with values which exceed 0.5 are considered to be
highly influential. If only one species is influential for a given source, it
is extremely important that its ambient and source measurements are accurate
and precise. The effects of influential species on the similarity/uncertainty
clusters and source contribution estimates can be determined by adding or
dropping this species as a fitting species.
TRANSPOSE OF SENSITIVITY MATRIX
SPECIES
NA
HG
AL
SI
CL
K
CA
TI
V
CR
HN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
N03
MAR IN
.99
.21
-.17
-.02
1.00
-.04
.28
.00
.23
-.10
-.02
-.08
.17
-.12
-.02
.01
.00
-.04
-.01
-.66
-.00
UDUST
.01
-.05
-.03
1.00
-.02
.05
.02
.49
-.04
-.07
-.07
-.12
-.06
-.19
-.09
-.02
-.04
.09
.03
.01
.01
AUTPB
-.03
-.04
-.10
-.05
.04
-.03
.03
-.07
-.06
-.03
-.10
-.05
-.06
.09
.07
.55
1.00
.61
.16
-.04
.03
SOURCE
RDOIL
-.09
-.04
-.11
-.05
.12
-.05
.12
.08
1.00
-.12
-.04
-.14
.87
.02
.09
-.05
-.06
.04
.17
.04
.06
CODE
KRAFT
.14
-.07
-.05
,04
-.36
.11
-.19
-.03
-.40
.13
-.07
.07
-.32
.05
-.02
.03
.00
.01
-.07
1.00
-.04
ALPRO
.01
.13
1.00
-.14
-.04
-.01
-.03
-.02
-.04
-.02
-.03
-.06
-.02
.17
-.05
-.04
-.07
.02
.20
-.04
.04
STEEL
-.06
.18
-.14
-.19
.01
-.22
.39
-.07
-.19
.33
-.05
1.00
-.09
.67
.31
-.07
-.08
-.22
-.16
-.16
-.19
FERMN
.01
-.08
.01
.02
-.02
.46
-.06
.02
.04
-.12
1.00
-.31
-.00
-.18
.10
-.01
-.05
.11
.09
-.01
.35
Strike enter to continue (ENTER)
1 Change Fitting Species
15 Exit
Type the line number to select: 13
Command 13 provides a visual display of CMB information. Bar charts of
calculated and measured chemical concentrations, bar charts of source
34
-------
profiles, and pie charts of source contribution estimates can be produced with
this command. Upon typing command 13, a graphics menu appears.
1 Graph elemental concentrations
2 Graph source profiles
3 Graph source contributions
4 Graph PM10
5 Exit graph menu
Type the line number to select or deselect 1
The plot shown in Figure 2 appears on the computer display. Note that
this plot displays both the calculated and measured concentrations of each
chemical species. Thick lines correspond to the measured concentrations,
while thin lines and "*" indicate calculated and fitting elements,
respectively. Typing ENTER returns the prompt
Hardcopy? Y or N
Y
Entering 'Y' results in the plot written to a file in HPGL language.
This is the language which drives Hewlett Packard Plotters and can be read by
many graphics, word processing, and desktop publishing programs. Figure 2 was
integrated into this manual from the CMBPLOT file using WordPerfect 5.0.
Plot file name cmbplot.OOl
After this prompt, the graphics menu reappears
1 Graph elemental concentrations
2 Graph source profiles
3 Graph source contributions
4 Graph PM10
5 Exit graph menu
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu: 2
Entering selection 2 displays the source profiles available for
plotting.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
MAR IN
CDUST
UDUST
AUTPB
RDOIL
VBRN1
VBRN2
KRAFT
SULFT
HOGFU
ALPRO
STEEL
FERMN
CARBO
GLASS
CAR8F
S04
35
-------
SITE PACS1 DATE 770813 DURATION 24 START HOUR 0 SIZE FINE
100, 000
10.000
C 1.000
0
N
C
E .100
N
T
R
A .010
T
I
0
N .001
_
_
—
*•
—
*
dfc
*
-V-
*
r*-
_
"*
_£
i —
•^
=*=
*=j
*
*
*-
*
*
£
TFNMASSCKCTVCMFNCZBPOESN
0 AGLI L At RNEIUNRBCCOO
T 43
Figure 2. Example of bar chart from CMB graphics menu.
18
19
20
18
19
20
N03
OC
EC
Toggle selection. Up or Down, Carriage return to exit: 4
Selecting source 4 causes an asterisk to appear beside this source
profile mnemonic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
MAR IN
CDUST
UDUST
AUTPB
RDOIL
VBRN1
VBRN2
KRAFT
SULFT
H06FU
ALPRO
STEEL
FERHN
CARBO
GLASS
CARBF
36
-------
17 17
18 18
19 19
20 20
S04
N03
OC
EC
Toggle selection. Up or Down, Carriage return to exit: (ENTER)
ENTER causes the display similar to that of Figure 2 to be plotted for
source profiles. Another ENTER yields
Hardcopy? Y or N
N
1
2
3
4
5
Graph elemental concentrations
Graph source profiles
Graph source contributions
Graph PM10
Exit graph menu
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 3
3.
Command 3 in the graphics menu presents a pie chart as shown in Figure
SITE PACS1
DATE 7-70813 DURATION 24 START HOUR 0 SIZE FINE
SOURCE AND « OF TOTAL MASS
RDOIL
14X
ALPRO
AUTPB
FERMN
MAR IN
15X
15X
KRAFT 5%
STEEL 11*
UDUST 12X
TOTAL OF NON-NEGATIVE SOURCES = 79.17073 = 99* OF MEASURED MASS
Figure 3. Example of pie chart from CMB graphics menu.
37
-------
If both FINE and COARS particle-size ranges are included and have been
apportioned, invoking Graphics Command 4 will plot a pie chart of the sum of
these two fractions. Invoking Graphics Command 5 returns control to the main
menu.
1 Change Fitting Species
15 Exit'
Type the line number to select: 6
It is often convenient to perform many CMB calculations at once and to
record the data on disk without interaction. To accomplish this, we invoke
Command 6, Autofit. Autofit allows a single selection of fitting species and
profiles to be applied to a selected list of samples without operator
intervention. This feature is especially useful for model simulation testing
and screening purposes. Autofit is equivalent to invoking commands 5, 9, 12,
and 4 in succession until the source contributions for all selected samples
have been calculated. Autofit first preselects all samples and then enters
Select Samples menu to allow deselection of any samples not to be included in
the selected list.
1 PACSl 08/13/77 24 0 COARS
2 PACSl 08/13/77 24 0 FINE
3 PACS2 01/24/78 24 0 COARS
4 PACS2 01/24/78 24 0 FINE
5 PACS3 08/07/77 24 0 COARS
6 PACS3 08/07/77 24 0 FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: (ENTER)
Exiting the Select Samples menu starts Autofit. There will be much disk
activity but no screen displays until Autofit has completed its analysis. All
of the CMB analyses may be printed from the OUPORT.DT1 and OUPORT.DT2 files.
1 Change Fitting Species
15 Exit'
Type the line number to select: 10
Command 10 creates a summary table of source apportionments for the
series and calculates the average and standard deviation of the source
contributions for each source profile used in the series.
OUTPUT WILL GO TO HAROCOPY.
DO YOU WANT IT DISPLAYED AT YOUR TERMINAL INSTEAD?
Y
The following output display can be directed to the OUPORT.DT1 file by
entering 'N' instead of 'Y' at the prompt above.
38
-------
FINE COARS TOTAL
CMB SITE DATE SOURCE (U6/M3) (UG/M3) (U6/M3)
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
08/13/77 MARIN
01/24/78 MARIN
08/07/77 MARIN
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 UDUST
01/24/78 UDUST
08/07/77 UDUST
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 AUTPB
01/24/78 AUTPB
08/07/77 AUTPB
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 RDOIL
01/24/78 RDOIL
08/07/77 RDOIL
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 KRAFT
01/24/78 KRAFT
08/07/77 KRAFT
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 ALPRO
01/24/78 ALPRO
-- OB/07/77 ALPRO
AVERAGE
(STD. DEV.)
DATE SOURCE
08/13/77 STEEL
12.39
-3.42
20.40
9.79
12.12
FINE
(UG/M3)
9.59
1.17
79.18
29.98
42.82
FINE
(UG/M3)
10.08
17.07
23.93
17.03
6.92
FINE
(UG/M3)
11.06
.74
14.01
8.60
6.97
FINE
(UG/M3)
4.69
14.23
-3.53
5.13
8.89
FINE
(UG/M3)
10.60
-.44
-1.52
2.88
6.71
FINE
(UG/M3)
8.67
10.60
.31
15.51
8.81
7.76
COARS
(UG/M3)
9.60
54.03
61.89
41.84
28.20
COARS
(UG/M3)
9.09
3.21
35.26
15.85
17.06
COARS
(UG/M3)
9.71
.19
12.45
7.45
6.43
COARS
(UG/M3)
12.33
.28
9.42
7.34
6.29
COARS
(UG/M3)
11.10
4.19
4.46
6.58
3.91
COARS
(UG/M3)
8.16
22.99
-3.12
35.91
18.60
19.88
TOTAL
(UG/M3)
19.19
55.20
141.07
71.82
62.62
TOTAL
(UG/M3)
19.17
20.28
59.19
32.88
22.79
TOTAL
(UG/M3)
20.77
.93
26.46
16.06
13.40
TOTAL
(UG/M3)
17.02
14.51
5.89
12.47
5.84
TOTAL
(UG/M3)
21.70
3.75
2.94
9.46
10.61
TOTAL
(UG/M3)
16.83
39
-------
PACS2
PACS3
01/24/78 STEEL
08/07/77 STEEL
AVERAGE
(STD. DEV.)
-1.00
-.02
2.55
5.33
.56
-2.21
2.17
5.37
-.45
-2.23
4.72
10.53
FINE COARS TOTAL
CMB SITE DATE SOURCE (UG/M3) (UG/M3) (UG/M3)
PACS1
PACS2
PACS3
08/13/77
01/24/78
OB/07/77
FERMN
FERMN
FERMN
AVERAGE
(STD. DEV.
)
11.
4,
6
.88
.62
.09
.20
.66
9
-
1
3
5
.87
.11
.41
.72
.38
21
1
7
11
.75
.51
.50
.92
.99
Strike enter to continue (ENTER)
1 Change Fitting Species
15 Exit
Type the line number to select: 15
All of the CMB menu commands have been illustrated in this tutorial, and
the final Command 15 exits from the CMB modeling session.
c>
40
-------
SECTION 4
INPUT AND OUTPUT DATA FILES
This section describes the structure of CMB7 input and output files and
methods of generating these files.
4.1 CMB7 INPUT AND OUTPUT FILE DESCRIPTIONS
All CMB input and output files are named in the form PPXXXXXX.SSS.
• The PP prefix indicates the contents of the file.
• The XXXXXX is replaced by a user to identify the specific combination
of data which he has placed in the file.
• The SSS extender indicates the format in which the data have been
arranged.
For example, the filename INSNW7E1.IN7 could remind the user that this
is the filename input file (IN) from the State of Nevada PM10 network (SN) for
winter (W) of 1987 (7) on EPA sampling days (E) at site 1 (1) in CMB7 text
format (.IN7). The recommended prefixes and suffixes are identified in the
descriptions of each individual file. A positioning line is situated above
each example of input files to show where each field begins and ends. Some
files are fixed format, which means that each field must begin and end on a
certain column. Other fields are blank delimited, which means data fields are
distinguished by one or more blank spaces between each field. Input data
files described by U.S. EPA (1987a) in CMB 6.0 can be used in CMB7 with the
following minor change:
• Insert four blank spaces before the element or source ID number in
the POXXXXXX.DAT and SOXXXXXX.DAT.
This change can be made in a text editor (or a word processing program)
and is needed because CMB7 views these numbers as six characters instead of
41
-------
two digits. Other than this change, the descriptions of CMB 6.0 input data
files are fully described by U.S. EPA (1987a) and those descriptions will be
briefly documented in Section 4.3. The CMB 6.0 formats should only be used if
they already exist. All new CMB input files should be constructed according
to the formats described in this document. CMB 6.0 formats will not be
«
supported in subsequent versions of the CMB software.
4.1.1 Input Filename File: INXXXXXX.IN7
This fixed format file contains a list of the names of other CMB7 input
data files. This filename, which is entered in response to the first prompt
when the CMB software is started, consists of seven lines as shown below.
Lines 1 through 3 contain any space filler such as 'XXXXXX'. Lines 4 through 7
contain the names of the files which are described in the following sub-
sections. INPORT.IN7 is an example of this file structure used in CMB7.
File: INPORT.IN7
0 1 2
12345678901234567890
Begin File:
XXXXXX
XXXXXX
XXXXXX
SOPORT.IN7
POPORT.IN7
ADPORT.TXT
PRPORT.TXT
End file.
Each filename can be up to eight characters in length with up to a
three-character extender. Each filename in this input filename file must be
left-justified and occupy Columns 1-12, Rows 4 through 7. The purpose of this
file is to save the effort of keying in the input filename individually. If
an INXXXXXX.IN7 filename is not entered at the prompt, CMB7 will request the
names of individual data input filenames.
4.1.2 Source Profile Selection File: SOXXXXXX.IN7
The most commonly used source profiles are usually identified in initial
CMB modeling sessions. It is often convenient to designate these profiles as
42
-------
defaults which will be selected each time the model is used. These default
profiles can be designated in the SOXXXXXX.IN7 file. Each line of this file
contains a six-character source code number in Columns 1 to 5 and an eight-
character source profile name in Columns 9-16. Source codes need not be
entered in numerical order, though this arrangement is helpful for file
editing and documentation. If the source is to be included as an initial
fitting source (as defined in Section 3.1.2), an asterisk is placed in Column
19. All other columns on each line are usually blank, but Columns 21-80 can
be used for comments or source documentation which can be viewed during screen
editing or when the file is printed. The maximum number of sources which can
be used in CMB7 depends on several variables, as described in Appendix A.
Fifty to 100 source profiles can be easily accommodated in most cases.
SOPORT.IN7 is an example of the source profile selection file.
*
*
*
File: SOPORT.IN7
1 2
12345678901234567890
Example:
Begin File:
1 MARIN
2 CDUST
3 UDUST
4 AUTPB
5 RDOIL
6 VBRN1
7 VBRN2
8 KRAFT
9 SULFT
10 HOGFU
11 ALPRO
12 STEEL
13 FERMN
14 CARBO
15 GLASS
16 CARBF
17 S04
18 N03
19 OC
20 EC
End File.
Optional Comments
Marine Aerosol
Continental Dust
Urban Dust—Better than CDUST in urban cases.
Leaded Auto Exhaust
Residual Oil Combustion
Vegetative Burning Profile 1
Vegetative Burning Profile 2
Kraft Paper Mill
Sulfite Paper Mill
Hogged Fuel Boiler
Aluminum Production
Steel Blast Furnace
Ferromanganese Furnace
Carborundum Furnace
Glass Furnace
Carborundum Furnace
Single Constituent Sulfate
Single Constituent Nitrate
Single Constituent Organic Carbon
Single Constituent Elemental Carbon
43
-------
As noted in the example above, comments which describe the meaning of
the source profile identifier can be entered in each record of this file after
the 20th space to further describe the source mnemonic or the reason it was
selected as a default value.
4.1.2 Species Selection File: POXXXXXX.IN7
The most commonly used fitting species are usually identified by initial
CMB modeling. These can be selected in POXXXXXX.IN7 as defaults which will be
selected as fitting species at model startup. This feature alleviates the
need to select these species individually every time the model is used.
This file structure is similar to that of the Source Names file. A six-
character species code is placed in Columns 1 to 6 and an eight-character
species name is placed in Columns 9 to 16. An asterisk in Column 19
designates that species as a fitting species. When Column 19 is blank, the
species is a floating species. The code '01' is reserved for particulate mass
and CMB7 assigns the species name 'TOT' to this code. In order to maintain
consistency with species coding in the EPA Source Profile Library, atomic
numbers of elements are recommended to be used for elemental species. The
total number of species which can be accommodated by CMB7 depends on several
variables, as described in Appendix A. Fifty to 100 different species can be
easily accommodated in most applications.
POPORT.IN7 is an example of the chemical species file.
File: POPORT.IN7
1 2
12345678901234567890 Optional Comments
Begin File:
Cl TOT Total Mass by Gravimetry
C9 F Water-soluble Fluoride by 1C
Cll NA * Sodium by Short-lived Neutron Activation
C12 MG * Magnesium by INAA
C13 AL * Aluminum by X-ray Fluorescence
C14 SI * Silicon by XRF--Marker for dust
C16 S Sulfur by X-ray Fluorescence
C17 CL * Chlorine by X-ray Fluorescence
C19 K * Potassium by X-ray Fluorescence
C20 CA * Calcium by X-ray Fluorescence
C22 TI * Titanium by X-ray Fluorescence
C23 V * Vanadium by X-ray Fluorescence
C24 CR * Chromium by X-ray Fluorescence
44
-------
C25 MN * Manganese by X-ray Fluorescence
C26 FE * Iron by X-ray Fluorescence
C28 NI * Nickel by X-ray Fluorescence
C29 CU * Copper by X-ray Fluorescence
C30 ZN * Zinc by X-ray Fluorescence
C35 BR * Bromine by X-ray Fluorescence
C82 PB * Lead by X-ray Fluorescence
C201 OC * Organic Carbon by TOR
C202 EC * Elemental Carbon by TOR
C203 S04 * Sulfate by An ion Chromatography
C204 N03 * Nitrate by Anion Chromatography
End File.
Text comments can be added to this file after the 20th character to
describe the species represented by the mnemonic. This is especially useful
when the same species has been measured by different methods.
4.1.3 Ambient Data Input File: ADXXXXXX.TXT
The first record of the ADXXXXXX.TXT file contains identifiers for each
field. These identifiers can be up to six alphanumeric characters in length
and are separated by a blank. CMB7 software uses these identifiers in the
order given in-this file to identify species concentrations and their
precisions in screen and output displays.
The ADXXXXXX.TXT file contains records with the following entries:
Field 1: Site ID (up to 12 characters)
Field 2: Sampling date (up to 8 characters)
Field 3: Sample duration (up to 2 characters)
Field 4: Sample start hour (up to 2 characters)
Field 5: Particle size fraction (up to 5 characters)
Field 6: Mass concentration (any number of characters in integer,
floating point, or exponential format)
Field 7: Precision of mass concentration (same format as Field 6)
Field 8+2n:Concentrations-of chemical species (same format as Field 6),
where n=0, 1, 2,
Field 9+2n: Precisions of species concentrations (same format as Field 6),
where n = 0, 1, 2,
Data records containing the blank delimited fields (i.e., data fields
are distinguished by a blank between each field) specified above follow this
first record. The total number of records which can be included depends on
the number of species, number of sources, and size of the computer memory. In
45
-------
most cases, one hundred or more records can be included in a single model
input file.
Up to four different size fraction identifiers may be used, and the user
can select mnemonics which suit his purposes. The size fraction names FINE
and COARS are reserved for the PM2.S and coarse particle (PM10-PM2-5) size
fractions which are most commonly measured in PM10 source assessment studies.
When these mnemonics are used, a separate output display is produced which
sums the FINE and COARS source contribution estimates to provide the estimates
for PM10.
Missing values for chemical species are indicated by placing a -99. in
the species concentration and precision fields. A species for which the value
is missing cannot be used as a fitting species. Precisions which exceed zero
must be assigned to all chemical concentrations which will be used as fitting
species. CMB7 will return an error message when it finds zero or negative
precisions.
ADXXXXXX.TXT files can be generated from dBASE III+ ambient data files
by first modifying the structure to contain the fields identified above, then
running the CMBIN7.PRG program (instructions in Section 4.2). Input files can
also be generated from spreadsheet files by arranging columns in the specified
order and using the spreadsheet's "print to disk" or "text output" options.
These files can also be created in a text editor which allows the editing of
very long records (VEDIT is such an editor). CMB7 file structures lend
themselves to data which are already in data base or spreadsheet formats,
however, and CMB 6.0 formats lend themselves more readily to data entry via
text editor.
The ADPORT.TXT file is an example input ambient data file in CMB7
format.
4.1.4 Source Profile Data Input File: PRXXXXXX.TXT
The PRXXXXXX.TXT file contains records with the following entries:
Field 1: Profile number or source code (up to six characters)
Field 2: Source mnemonic (up to eight characters)
Field 3: Particle size fraction (up to five characters)
Field 4+2n: Fraction of species in primary mass of source emissions
(floating point or exponential format), where n = 0, 1, 2, ...
46
-------
Field 5+2n: Variability of fraction of species in primary mass of source
emissions (same format as Field 4), where n = 0, 1, 2,
The first record of the PRXXXXXX.TXT file contains identifiers for each
field. These identifiers can be up to six alphanumeric characters in length,
are blank delimited, and must correspond to the identifiers used in the
ADXXXXXX.TXT file. The order of the chemical species in the PRXXXXXX.TXT file
must be the same as the order specified in the ADXXXXXX.TXT and POXXXXXX.IN7
files.
Data records containing the blank delimited fields specified above
follow this first record. The total number of records which can be included
depends on the number of species, number of sources, and size of the computer
memory. In most cases, one hundred or more records can be included in a
single model input file.
Up to four different size fraction identifiers may be used, and the user
can select mnemonics which suit his purposes. These identifiers must
correspond to those used in the ADXXXXXX.TXT file.
Missing values for chemical species in source files are replaced by a
best estimate with a large variability. Default values of 0 for the fraction
and .0001 for the variability are often chosen for species which are expected
to be present in small concentrations. This indicates that the species is
present in source emissions at a concentration less than .01%. A smaller
value may be appropriate for certain source-types and species. A precision
value which exceeds zero must be entered for all fitting species. CMB7 will
return an error message when it detects precisions which are less than or
equal to zero.
PRXXXXXX.TXT files can be generated from single record dBASE III+ source
profile files by first modifying the structure to contain the fields
identified above, then running the CMBIN7.PRG program (instructions in Section
4.2). Input files can also be generated from spreadsheet files by arranging
columns in the specified order and using the spreadsheet's "print to disk" or
"text output" options. These files can also be created in a text editor which
allows the editing of very long records (VEDIT is such an editor). CMB7 file
structures lend themselves to data which are already in data base or
47
-------
spreadsheet formats, however, and CMB 6.0 formats lend themselves more readily
to data entry via text editor.
PRXXXXXX.DBF files can be generated from the EPA source composition
library dBASE III+ file structures using the PROFIN.PRG program (instructions
are in Section 4.2). The PRPORT.TXT file is an example of an input source
profile data file in CMB7 format.
4.1.5 Hardcopv Output File: OUXXXXXX.DT1
The hardcopy output file presents the source contribution estimates,
standard errors, model performance measures, and measured and calculated
chemical species on a one-page display for each sample. The written display
is identical to that which appears on the screen during an interactive
modeling session. This output can be printed to hardcopy and bound in a
modeling report. The file format is in ASCII text mode and is self-
explanatory. Examples of output are given in Section 3.
4.1.6 Data Analysis Output File: OUXXXXXX.DT2
The data analysis output file records the contribution of each source-
type to each chemical species in a single data record. Sample identifiers and
model performance measures are also included in each record. All fields are
blank-delimited. The file structure is:
Field 1: Species identifier
Field 2: Fitting flag; a '*' indicates that the species was a fitting
species, while a '_' indicates that it was not
Field 3: Sampling site identifier
Field 4: Sampling date
Field 5: Sample start hour
Field 6: Sample duration
Field 7: Particle size fraction
Field 8: Measured species concentration
Field 9: Precision of measured species concentration
Field 10: R square value
Field 11: Chi square value
Field 12: Percent of measured mass
Field 13+2n: Source contribution estimate, n = 0, 1, 2,
Field 14+2n: Standard error of source contribution estimate, n = 0,
48
-------
Fields 1 and 3 through 9 repeat model input data. Fields 2 and 10
through 12 provide information about the CMB calculation. The remaining
fields correspond to each source profile in the PRXXXXXX.TXT file and contain
the source contribution estimates and standard errors for these sources. A
value of -99. is recorded when a profile was not used in the calculation.
The first record in this output file contains blank delimited field
identifiers. All subsequent records contain data. Fields 13+2n and 14+2n are
labeled with source codes and source contribution uncertainty columns are
labeled with source names.
4.1.7 Graphics Output File: CMBPLOT.XXX
In addition to CMBOUT.DT1 and CMBOUT.DT2, CMB7 also gives the user the
option to produce Hewlett-Packard Graphics Language (HPGL) plot files of the
screen graphic displays. Each plot is placed into a separate file of the same
name with the extension incremented by one (e.g., CMBPLOT.001, CMBPLOT.002,
etc.).
These plots can be read directly into many graphics and desktop
publishing packages. Publication-quality plots can also be produced from
these files on HP 7470 and 7475 plotters, or HP LaserJet II printers, using
software or hardware previously described in Section 2.1.
4.2 CREATING DATA INPUT FILES
There are three common methods of creating the CMB input files:
manually entering the data in the correct format using a text editor or word
processing program; editing existing input files with a text editor or word
processing program; or transferring files from computerized data bases.
A text editor or word processor in text mode can be used to type entire
input files. It is best to bring the example files into the editor, then
insert the new values in the same locations..as the existing values by using
the editor in TYPEOVER mode. All spaces between fields should be entered with
the space bar; tabs should not be set. Each line should be terminated with
»
the ENTER key rather than using the wraparound feature present in many
editors. No blank lines at the end of the file should be present. Completed
files should be saved with an appropriate filename. EDLIN, the line editor
49
-------
supplied with MS/DOS or PC/DOS, may be used, but it is not convenient since it
is not screen oriented.
If input data files have been prepared for other applications (e.g.,
source profiles may be common to several different data sets), then these
files may be cut and pasted to produce the needed input data files. Because of
differences in individual editing programs, the user is advised to consult the
manual for the editing program to be used for directions on opening a copy of
the existing file, deleting and adding material, saving the changes, and
renaming the file.
Many source profile and ambient data sets are available in data base
management formats. Selections of data, field names, and data structure can
be easily made by the data base software.
4.2,1 Conversion of EPA Source Data Base Files
The dBASE III+ program PROFIN.PRG selects source profiles and species
from the EPA source composition library data base and places these into a
dBaseIII+ file which can then be formatted into an ASCII input file for CMB7.
An example of the EPA source composition library is included in the
PROFILE.DBF example data base. The following tutorial using this program
requires the dBaseIII+ or dBase IV data management software.
The dBase software must reside in the same directory as the CMB files,
or a path to the dBase software must be established with the DOS PATH command,
to successfully implement file conversions.
First, start dBase and obtain the dot prompt
008ASE
.00 PROFIN
ENTER THE NAME OF THE ELEMENT NAME FILE: POPORT.OAT
ENTER THE NAME OF THE OUTPUT DBASE PROFILE FILE: PRPORT1
ENTER THE NAME OF THE EPA SOURCE PROFILE DBASE FILE: PROFILE
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 1
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: Z
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 3
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 4
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 5
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 6
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 7
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 8
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 9 •
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 10
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 11
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 12
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 13
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 14
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 15
50
-------
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 16
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 17
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 18
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 19
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: 20
ENTER THE PROFILE NUMBER OR A CARRIAGE RETURN TO EXIT: (ENTER)
USE PRPORT1
COPY TO PRPORT FOR SIZE-'AO_25'.OR.SIZE-tA2_10'
USE PRPORT
REPLACE SIZE WITH 'FINE' FOR SIZE='AO_25f
REPLACE SIZE WITH 'COARS' FOR SIZE-'A2_10'
MODIFY STRUCTURE
Eliminate the fields 'Cl' and 'TOT' by placing the cursor on each one
with the up arrow or down arrow key and pressing CTRL-LJ. Press CTRL-W to save
these changes. Enter BROWSE and enter the desired source mnemonic in the name
column. Save these changes by typing CTRL-W. The remaining dBase file should
look like this:
CODE NAME SIZE C9 F ... C204 N03
1 MARIN FINE 0.00000 0.00000 ... 0.00000 0.00000
20 EC COARS 0.00000 0.00000 ... 0.00000 0.00000
The following commands will convert the dBase file (e.g., PRPORT.DBF),
which was prepared by the PROFIN.PRG program, into a text file (e.g.,
PRPORT.TXT).
.00 CMBIN7
NAME OF DBASE INPUT FILE: PRPORT
NAME OF CMB INPUT FILE: PRPORT
This PRPORT.TXT file is in CMB7 input data format. The ADPORT.TXT input
data file may be created by the DO CMBIN7 command and typing ADPORT instead of
PRPORT in response to the prompts.
4.3 CMB 6.0 FILES FORMATS
CMB 6.0 file formats are not as versatile or data-base oriented as CMB7
formats. They are still available for users who prefer them, however, and a
brief documentation is provided here. The files are:
• INXXXXXX.DAT. List of other input data filenames
• SOXXXXXX.DAT. List of source profile numbers and names
• POXXXXXX.DAT. List of species numbers and names
• FSXXXXXX.DAT. Fine particle source profiles
• CSXXXXXX.DAT. Coarse particle source profiles
• DAXXXXXX.DAT. Fine and coarse particle ambient
concentrations
51
-------
Each file except the DAXXXXXX.DAT file contains a single line format
(i.e., entries on sequential lines all have the same format). The formats for
the six files are presented in separate subsections below. The notations A,
I, and F refer to alphanumeric, integer, and floating-point formats,
respectively. A floating-point format designated F8.6, for example, indicates
that the field is eight characters wide and, if no decimal point appears in
the field it is assumed that six characters are to the right of the implied
decimal. Frequently, this decimal point is included in the data in a floating
point field; in that case, the actual decimal location overrides the implied
location in the "F" format. Data entries in the I format should always be
right justified (i.e., aligned to the right side of the field).
The INXXXXXX.DAT file contains the names of the five other filenames in
the following order:
File: INPORT.DAT
0 1 2
12345678901234567890
Begin File:
FSPORT.DAT
CSPORT.DAT
DAPORT.DAT
SOPORT.DAT
POPORT.DAT
End File.
SOXXXXXX.DAT and POXXXXXX.DAT are identical to SOXXXXXX.IN7 and
POXXXXXX.IN7 except that the species codes must be numeric, not alphanumeric.
The CSXXXXXX.DAT and FSXXXXXX.DAT files contain four required input
fields plus an unstructured comment area to describe the data recorded in
those fields. The format is:
COLUMN FORMAT CONTENTS
1-2 12 Source code number
3-6 14 Species code
9-16 F8.X Fraction of fine (or coarse) source
emissions from indicated species
19-26 F8.X Uncertainty of fraction
27-80 A60 Can be used for comments; usually to
identify size fraction, source name, and
species
52
-------
All source code numbers and species codes used in this file must be
listed in the SOXXXXXX.DAT and POXXXXXX.DAT files.
The DAXXXXXX.DAT file contains two line formats for each sample. The
header format (identified by the numbers 03 in columns 1 and 2) provides
information on the receptor site and date and is limited to a single line.
The second type of format (identified by the numbers 30 in columns 1 and 2)
records ambient concentrations for the above site and date, with a separate
line for each chemical species.
TYPE
Header
COLUMN
1-2
4-15
17-18
19-22
24-25
27-28
33-34
FORMAT1
12
A12
A2
A4
12
12
12
35-80
Concen-
tration
1-2
4-15
17-18
19-22
24-25
27-28
31-34
37-45
48-56
59-67
70-78
12
A12
A2
A4
12
12
14
F9.X
F9.X
F9.X
F9.X
CONTENTS
'03'
Receptor identification
Year YY
Month and day, MMDD
Duration of sample, hours
Starting hour of sample
Size fractions on next lines; 12 = fine and
coarse; 13 = fine and total
Must be blank
'30'
Receptor identification
Year, YY
Month and day, MMDD
Duration of sample, hours
Starting hour of sample
Species code
Concentration of fine fraction
Standard error of fine fraction (measurement
uncertainty)
Concentration of coarse or total fraction
Standard error of coarse or total fraction
(measurement uncertainty)
I - Integer (12; 0 through 99)
A - Alphanumeric (A12; up to 12 characters, A-Z, 0-9, any keyboard symbol)
F - Floating point (F9.X; up to 8 integers and a decimal point). Note: It
is very helpful in reading the file to have the decimal points aligned
vertically. A decimal point must be included in the field.
53
-------
An unlimited number of ambient samples can be included in this file.
Species codes must be the same in all files. Measurement precisions which
exceed zero must be assigned to the ambient concentration data.
If only PHjo data are available, all fine particle concentrations should
be set to zero and the "fine and total" designation (13) should be entered in
columns 33-34 of the header record. CMB analyses should be performed only on
the total size fraction.
54
-------
SECTION 5
CMB PERFORMANCE MEASURES
This section describes the different performance measures which are used
to evaluate the validity of source contribution estimates. Greater detail on
the use of the performance measures is presented by U.S. EPA (1987b). The
performance measures are presented in three separate displays when Commands 5
or 7 are invoked: 1) the source contribution display; 2) the uncertainty
similarity cluster display; and 3) the species concentration display. Each of
these displays is discussed below. The equations used to calculate these
measures are presented in Appendix A.
5.1 SOURCE CONTRIBUTION ESTIMATES DISPLAY
An example of a source contribution table display is shown below:
SOURCE CONTRIBUTION ESTIMATES - SITE: MCS1 DATE: 08/13/77 CMB7 8*331
SAMH.E DURATION 24 STMT HOUR 0 SIZE: COARS
R SQUARE .97 PERCENT MASS 100.6
CHI SQUMIE 1.44 OF 13
SOURCE
• TYPE SC£(U6/N3) STD ERR TSTAT
1
3
4
S
8
11
12
13
MARIN
UOUST
mm
ROOIL
KRAFT
ALPRO
STEEL
FERMN
10.6029
9.5985
9.0906
9.7127
12.32«S
11.0997
8.1587
9.8720
1.8240
1.2616
1.3961
1.6108
7.9102
2.2441
1.5Z39
1.6165
5.8131
7.6081
6.5112
6.0296
1.5583
4.94«2
5.3538
6.1072
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/ 160.04- 11.3
Source contribution estimates are the main output of the CMB model. The
sum of these concentrations approximates the total mass concentration.
Negative source contribution estimates are not physically meaningful, but they
can occur when a source profile is col linear with another profile or when the
source contribution is close to zero. Collinearity is usually identified in
the similarity/uncertainty cluster display. When the absolute value of a
positive or negative source contribution estimate is less than its standard
55
-------
error, the source contribution is undetectable. Two or three times the
standard error may be taken as an upper limit of the source contribution in
this case.
The standard errors reflect the precisions of the ambient data, the
source profiles, and the amount of collinearity among different profiles.
Standard errors should be reported with every source contribution estimate.
The standard error is a single standard deviation. There is about a 66%
probability that the true source contribution is within one standard error and
about a 95% probability that the true contribution is within two standard
errors of the source contribution estimate.
The T-statistic (TSTAT) is the ratio of the source contribution estimate
to the standard error. A TSTAT value less than 2.0 indicates that the source
contribution estimate is at or below a detection limit. Low TSTAT values for
several source contributions may be caused by collinearities among their
profiles; this will be indicated by the Similarity/Uncertainty Clusters.
The reduced chi square, degrees of freedom, R square, and percent mass
are goodness of fit measures for the least squares calculation.
The chi square is the weighted sum of squares of the differences between
the calculated and measured fitting species concentrations. The weighting is
inversely proportional to the squares of the precisions in the source profiles
and ambient data for each species. Ideally, there would be no difference
between calculated and measured species concentrations and chi square would
equal zero. A value less than 1 indicates a very good fit to the data, while
values between 1 and 2 are acceptable. Chi square values greater than 4
indicate that one or more species concentrations are not well explained by the
source contribution estimates. The degrees of freedom equal the number of
fitting species minus the number of fitting sources. The degree of freedom is
needed when statistical significance tests are applied to the reduced chi
square value.
The R square is the fraction of the variance in the measured
concentrations data which is explained by the variance in the calculated
species concentrations. It is determined by a linear regression of measured
versus model-calculated values for the fitting species. R square ranges from
0 to 1.0. The closer the value is to 1.0, the better the source contribution
56
-------
estimates explain the measured concentrations. When R square is less than
0.8, the source contribution estimates do not explain the observations very
well with the given source profiles and/or species.
Percent mass is the percent ratio of the sum of the model-calculated
source contribution estimates to the measured mass concentration. This ratio
should equal 100%, though values ranging from 80 to 120% are acceptable. If
the measured mass is very low (< 5 to 10 Mg/nr)» this ratio may be outside of
this range because the precision of the mass measurement is about 2 fig/vr (EPA
o
sampling requirements are 5 #g/m precision).
5.2 SIMIURITY/UNCERTAINTY CLUSTER DISPLAY
This display is shown below:
UNCERTAINTY/SIMILARITY CLUSTERS CMB7 89338 SUM OF CLUSTER SOURCES
1 8 17.078+- 4.241
1 5 8 28.139-t- 3.833
The first column contains the clusters, one cluster on each row. Each
cluster is identified by the code numbers associated with its source profile.
The clusters are formed when: 1) two or more source profiles in an eigenvector
derived from the singular value decomposition exceed 0,25 (these terms are
explained in Appendix A); and 2) the T-statistic for any one of these source-
types is less than or equal to 2.0. These uncertainty/similarity clusters are
caused by excessive similarity (collinearity) among the source profiles in the
cluster or by high uncertainties in the individual source profiles. The
standard errors associated with the source contribution estimates of one or
more sources identified in a cluster are usually very large, often too large
to allow an adequate separation of these source contributions to be made.
If collinearity is the cause of these excessive standard errors, then
the standard error of the sum of the source contributions for a cluster may be
smaller than the standard error of any single source contribution in the
cluster. The sum of source contributions and the standard error of the sum
are shown in the final column of this display. This sum may be more useful
than the individual source contribution estimates if the standard error of the
57
-------
sum is substantially lower than the standard errors of each source
contribution estimate. The sum does not allow differentiation among the
contribution estimates of the sources contained in the cluster.
Clusters will not appear if the two above-stated criteria are not met.
This absence of clusters means that the source contributions can be resolved
in the specific application. Since ambient data uncertainties, and relative
levels of source contributions vary from sample to sample, it is possible that
a given combination of profiles may appear in the clusters for one set of
ambient data, but not for another set. For this reason, it is impossible to
decide a priori that a set of profiles is collinear or not. The decision must
be made for each set of data and each set of profiles combined with those
data.
5.3 SPECIES CONCENTRATIONS DISPLAY
An example of the species concentration display is shown below:
SPECIES CONCENTRATIONS - SITE: PACS1
SMVU DURATION 24 STMT HOUR
R SDUME .90 PERCENT MASS
CHI SQUARE 4.85 OF
DATE: 08/13/77 CNB7 89338
0 SIZE: COMS
94.8
12
SPECIES 1 — «AS
Cl
C9
cn
C12
CIS
C14
C16
C17
C19
C20
C22
C23
C24
CZS
C26
C28
C29
CM
as
C82
C201
C202
C203
C204
TOT T 80.00000+-
F .734004-
NA
K
H
SI
S
CL
K
«
TI
v
CR
MN
FE
NI
CU
IX
8R
n
OC
EC
SO*
6.33000*-
1.48000*-
4.84000*-
3.27000*-
2.50000*-
4.67000*-
1.12000*-
1.52000*-
.14000*-
.27700*-
.00800*-
2.47000*-
5.41000*-
.77900*-
.05000*-
.21400*-
.52000+-
1.78000*-
10.10000+-
1.68000*-
8.10000*-
N03 1.13000+-
8.00000
.07300
.63300
.14800
.48400
.32700
.25000
.4*700
.11200
.15200
.01400
.02800
.00100
.24700
.54100
.07800
.00500
.02100
.05200
.17800
1.01000
.16800
.81300
.11300
•-CAIC
75.83980*-
.62178*-
S.5911S*-
1.15164+-
S.316044-
3.65624*-
2.28332*-
5.74261+-
1.90394*-
1.14509*-
.14879*-
.41650*-
.01272*-
2.56464*-
1.52394*-
.65911*-
.04401*-
.18275*-
.62465*-
2.28S33*-
8.11789*-
1.41114*-
7.95489*-
1.028Z7*-
RATIO C/W RATIO R/U
3.68599
.25*80
.56393
.70809
.67496
.19443
.33567
1.30243
.58130
.11261
.02506
.08825
.00580
.14741
.24114
.14252
.00719
.04540
.19393
.33450
1.36887
.40161
1.50170
.50149
.95*-
.8S+-
1.04*-
.78*-
1.10*-
I.I2*-
.91+-
1.23+-
1.70+-
.75*-
1.06*-
1.50+-
1.59+-
1.04*-
.28*-
.85+-
.88*-
.85*-
1.20*-
1.28*-
.80*-
.84*-
.98+-
.91+-
.11
.36
.14
.48
.18
.13
.16
.30
.55
.11
.21
.35
.75
.12
.05
.20
.17
.23
.39
.23
.16
.25
.21
.45
-.5
-.4
.3
-.5
.6
1.0
-.5
.8
1.3
-2.0
.3
1.5
.8
.3
-6.6
-.7
-.7
-.6
.5
1.3
-1.2
-.6
-.1
-.2
This display shows how well the individual ambient concentrations are
reproduced by the source contribution estimates. This display offers clues
concerning which sources might be missing or which ones do not belong in the
58
-------
calculation. Fitting species are marked with an asterisk in the column
labeled T.
The column labeled RATIO R/U contains the ratio of the signed difference
between the calculated and measured concentrations (the residual) divided by
the uncertainty of that residual (square root of the sum of the squares of the
uncertainty in the calculated and measured concentrations). The R/U ratio
specifies the number of uncertainty intervals by which the calculated and
measured concentrations differ. When the absolute value of the R/U ratio
exceeds 2, the residual is significant. If it is positive, then one or more
of the profiles is contributing too much to that species. If it is negative,
then there is an insufficient contribution to that species and a source may be
missing. The sum of the squared R/U for fitting species divided by the
degrees of freedom yields the chi square. The highest R/U values for fitting
species are the cause of high chi square values.
The column entitled RATIO C/M shows the ratio of calculated to measured
concentration and the standard error of that ratio for every chemical species
with measured data. The ratios should be near 1.00 if the model has
accurately predicted the measured concentrations.
5.4 ADDITIONAL DIAGNOSTICS
Command 8 (Present Source Contributions) of the main menu produces a
table that shows the fraction of each species' calculated ambient
concentration contributed by each source in the fit. An example of this
display is shown below:
CM.C SPECIES(PER SOURCE)
INDIVIDUAL RATIO .
MEM SKCIESfAU SOURCES)
SOURCE CODE
SPECIES NMIN UOUST AUTP8 RDOIL KRAFT ALPRO STEEL FERM
TOT .132 .121 .113 .102 .179 .138 .102 .124
F .000 .001 .000 .006 .000 .634 .000 .039
NA .667 .027 .000 .045 .120 .042 .016 .049
NG .343 .101 .000 .000 .000 .202 .359 .000
M. .000 .132 .021 .009 .DM .716 .011 .013
SI .000 .821 .023 .024 .006 .003 .125 .030
S .139 .000 .014 .436 .189 .000 .064 .067
CL .905 .000 .058 .000 .089 .028 .032 .009
K .132 .089 .006 .020 .051 .000 .067 .929
CA .097 .191 .074 .085 .034 .059 .333 .085
TI .000 .698 .000 .064 .000 .060 .117 .033
V .000 .009 .000 1.018 .000 .016 .018 .009
CR .000 .544 .000 .482 8.569 .221 21.466 .520
m .000 .004 .000 .002 .003 .000 .288 .694
FE .000 .102 .035 .045 .049 .008 .483 .038
III .000 .000 .002 .564 .040 .0)0 .073 .000
59
-------
u
ZN
SR
ft
DC
EC
504
.000
.000
.Ml
.000
.000
.000
.130
.058
.050
.001
.020
.032
.089
.001
.132
.148
.870
1.017
.448
.205
.015
.123
.153
.002
.005
.057
.151
.487
.171
.000
.015
.000
.223
.153
.201
.310
.005
.026
.000
.000
.105
.023
.458
.458
.000
.035
.000
.000
.025
.071
.269
.031
.003
.088
.089
.051
MO .000 .002 .073 .047 .000 .000 .000 .500
The sources which are major contributors to each species can be
determined by examining this display.
Another diagnostic is the transpose of the normalized modified pseudo-
inverse matrix (MPIN). An example display of this diagnostic is shown below:
TRANSPOSE OF SENSITIVITY HATSIX
SOURCE CODE
SPECIES NMIN UOUST MITPB RDOIL KRAFT ALPRO STEEL FERMI
M
IK
M.
SI
a
K
CA
TI
V
CD
m
FE
U
ZN
BR
ra
ac
EC
S04
N03
1.00
.27
-.02
-.04
.78
.01
.19
-.01
.10
-.17
-.03
-.04
-.26
.09
.09
.17
-.42
-.19
-.18
-.01
.01
-.04
.02
1.00
-.03
.07
.OS
.34
-.04
-.06
-.06
-.12
-.1!
-.11
-.OZ
-.04
.07
.06
.01
.02
-.04
.02
.01
-.03
.07
-.03
.06
-.02
.06
-.12
-.05
-.05
-.06
.14
.49
1.00
.23
.03
-.16
.03
-.06
.00
.00
-.02
-.02
-.02
.08
.03
1.00
-.15
-.04
-.06
-.09
.18
.02
• .06
-.21
.01
.42
.03
.10
-.23
-.18
.03
-.19
.09
-.20
-.03
-.56
.42
-.10
.00
.49
-.35
-.19
-.45
1.00
.40
.62
-.02
-.01
.10
1.00
-.12
-.01
.00
.01
.01
.02
-.08
-.00
-.12
-25
-.04
.01
-.00
-.11
.04
-.07
.01
-.10
.27
-.22
-.20
.05
-.30
.54
-.07
-.06
.29
-.03
1. 00
.40
.50
-.04
.01
-.38
-.22
-.29
-.16
.03
-.12
.09
.04
-.05
.52
-.15
.02
-.01
-.14
1.00
-.38
-.16
.02
.01
-.04
,14
.09
.07
.27
This matrix indicates the degree of influence each species concentration
has on the contribution and standard of error of the corresponding source
category. MPIN is normalized such that it takes on values from -1 to 1.
Species with MPIN absolute values of 1 to 0.5 are associated with influential
species. Noninfluential species have MPIN absolute values of 0.3 or less.
Species with absolute values between 0.3 and 0.5 are ambiguous but should
generally be considered noninfluential.
60
-------
SECTION 6
SOURCE AND RECEPTOR PARTICIPATE DATA BASES FOR THE CMB
One of the original objections to receptor modeling for PM18 source
assessment was that source profile and ambient data were not available for
their application. These objections are no longer valid, since a large number
of data bases of both source and receptor measurements have been acquired in
the United States for use in these models over the past decade. These data
bases are widely dispersed, however, and are not generally available for study
or evaluation. This situation is unfortunate because: 1) these existing
particulate data bases might alleviate the need to acquire new data bases; 2)
their potential for receptor model application and testing is untapped; 3)
they provide models for success and failure which can enhance the design of
new data acquisition projects; and 4) they provide a comprehensive view of
particulate levels, concentrations, and source contributions for major parts
of the United States. This section identifies several data bases which are
available and can be used in the future to address these goals.
6.1 DATA BASE REQUIREMENTS
It is not possible, nor even of value, to catalog every measurement of
particulate matter ever taken. Since many measurement programs have recently
been completed, several are in progress, and others are planned, any
compilation must be considered a snapshot at a particular time. Nevertheless,
such a snapshot is useful since it provides a starting point for evaluating
and using existing data. The information included here is useful today, and
it will serve as a starting point for future compilations of a similar nature.
The previous survey of this type was conducted for Total Suspended Particulate
(TSP) data bases by Lioy et al. (198C) nearly a decade ago.
The ideal particulate matter data base for source and receptor
measurements has the following characteristics:
61
-------
• A large number of chemically and size classified concentrations.
Mass, elements, ions, and carbon have been found to be the most
easily measured and useful species, while PMi0 and PM25 are the most
useful size ranges.
• Comprehensive coverage with respect to time, space, and, in the case
of source samples, operating conditions. Simultaneous receptor
samples taken at locations affected by different source-types are
useful in the verification of receptor model source apportionments.
Similarly, receptor samples taken in different seasons are affected
by different emissions sources and meteorological conditions. Source
samples need to represent the full range of profiles from a given
source category so that uncertainties can be estimated for input to
receptor models. Multivariate receptor models, such as multiple
linear regression or factor analysis, require more than approximately
70 receptor samples to comply with their assumptions.
• Documentation of measurement methods, locations, and sampling times.
Written records of the entire experimental program which acquired the
data base are essential to its extended use. In the case of source
characterization, this information should include the fuels,
operating cycle, type of facility, location, and time of test.
• Quality control and quality audits. Replicates, field blanks, and
independent verifications of field monitoring and laboratory
operations are needed to assure that the stated procedures were
actually complied with.
• Precision and accuracy estimates. State-of-the-art receptor modeling
treats measurement uncertainties as part of the input data and
returns uncertainties on source contribution estimates derived from
those inputs. The quality control and quality audit data should be
processed to quantify these uncertainties.
• Validation summaries or flags. Validation criteria should have been
applied to every sample, and the results of that validation should be
reported with the data.
• Availability in computerized formats. For research purposes, data
cannot be proprietary or secret. If it is not in some computerized
and documented form, preferably accessible by desktop computers, the
expense of putting it into such form usually outweighs the potential
benefits derived from examining the data.
6.2 DATA BASE SURVEY *
The data bases presented here generally meet the foregoing requirements,
though no single data base completely fulfills all of them. These data bases
62
-------
were identified via contacts with nearly 50 state and local agencies,
universities, and researchers. Nearly 100 reports and publications were
assembled and reviewed. The availability of the data in the public domain was
ascertained, and data sets which could not be released were excluded from
further consideration.
6.2.1 Source Characterization Data Bases
Table 1 summarizes the data bases which have been compiled for source
profiles. The EPA source composition library (U.S. EPA, 1988) is recommended
for Level I PM10 assessment. This library contains a large number of sources,
but its current contents for motor vehicles and residential wood combustion
are dated and do not reflect the compositions from modern vehicles, stoves,
and fuels. The more recent source libraries (Cooper et al., 1987; Ahuja et
al., 1989; Houck et al., 1989; Core et al., 1989; Watson et al., 1988, 1989)
contain profiles which are more applicable to data bases acquired today for
PMjo source assessment. The historical source libraries from the Portland
Aerosol Characterization Study (Watson, 1979) and the wide range of profiles
reported by Hopke (1985) are still of value because of their comprehensiveness
and applicability to receptor data taken in an earlier era. Sheffield and
Gordon (1986) present the most complete compilation of emissions
characteristics from coal- and oil-fired power plants, and this is an
excellent resource for studying pollution in areas with these source-types.
Most of the collections of source profiles listed in Table 1 contain soil and
road dust compositions, and it is unlikely that these profiles change over
long periods of time. The Pacific Northwest Source Composition Library (Core,
1989) is one of the first to acquire speciation of the organic carbon fraction
of source samples. The measurement of these additional species will allow
them to be evaluated in receptor model applications to particulate matter.
6.2.2 Receptor Measurement Data Bases
Several major ambient particulate studies are described in Table 2. The
notations are self-explanatory, and the reader is referred to the references
for additional information on each study. All of those listed in Table 2 are
63
-------
TABLE 1
SUMMARY OF CMB SOURCE PROFILES
Study (Reference)
EPA Receptor Model Source
Library (Core et al.,
1984; U.S. EPA, 1988)
South Coast Air Basin
Source Composition
Library (Cooper et al.,
1987)
California Air Resources
Board Source Library
(Ahuja et al., 1989;
Houck et al., 1989)
Pacific Northwest
Composition Library (Core
et al., 1989)
Receptor Model ing in
Environmental Chemistry
(Hopke, 1985)
Fine Particle Emissions
from Stationary and
Miscellaneous Sources in
the South Coast Air Basin
(Taback et al., 1979)
University of Maryland
Source Compositions
(Sheffield and Gordon,
1986)
Portland Aerosol
Characterization Study
(Watson, 1979)
Montana Source
Composition Library
(Houck et al., 1982;
1984; Pritchett et al.,
1985)
Size
Fractions
0 - 2.5 fan
2.5 - 10 fan
0 - 10 fan
0 - 30 fan
0 - 2.5 fan
2.5 - 10 im
0 - 10 /on
0
0
0
0
1.0 (an
2.5 fan
10 fan
30 fan
Q - 2.5 fan
2.5 - 10 i
0
0
0
2.5 fan
10 fan
30 fm
0 - 2.5 fan
0 - 2.5 fan
2.5 - 15 fan
0 - 30 fan
0 - 2.5 fan
0 - 30 fan
0 - 2.5 fan
2.5 - 15 fan
0 - 30 fan
Ma.ior Source Types
Geological, Motor Vehicles,
Vegetative Burning, Industrial
Geological, Motor Vehicles,
Vegetative Burning, Industrial
Geological, Diesel Trucks,
Vegetative Burning, Industrial
Geological, Motor Vehicles,
Vegetative Burning, Industrial
(Forest Products)
Geological, Motor Vehicles,
Vegetative Burning, Industrial
Industrial
Coal- and Oil-Fired Power
PI ants
Geological, Wood Burning,
Industrial
Geological. Motor Vehicles.
Wood Burning, Industrial
64
-------
TABLE 1 (continued)
SUMMARY OF CMB SOURCE PROFILES
Study (Reference)
Missoula City-County Air
Pollution Control
District Source Library
(Houck et al., 1987)
Wyoming Source
Composition Library
(Pritchett and Cooper,
1985a)
Alaska Source Composition
Library (Pritchett and
Cooper, 1985b)
Harvard Air Pollution
Health Effects Study
(Chow, 1985)
State of Nevada Air
Pollution Study (Watson
et al., 1988a)
SCENIC Denver (Watson et
al., 1988b)
Size
Fractions
0 - 2.5 fan
2.5 - 10 urn
0 - 2.5 fm
2.5 - 15 [
0 - 2.5 i
2.5 - 10
0 - 2.5 fan
0
0
0
0
2.5
10 i
2.5 fan
10 fm
Ma.ior Source Types
Geological, Motor Vehicles,
Wood Burning
Geological
Geological, Wood Burning
Geological
Geological, Motor Vehicles,
Wood Burning
Geological, Motor Vehicles,
Wood Burning, Industrial
(brewery, catalyst
cracker), Power Plant (coal
and gas)
65
-------
fairly major studies which were initiated for the purpose of applying receptor
models, and several of the references include the results of the modeling.
Hopke (1985) identifies a number of additional studies which are more short-
term than the applications orientation of the data bases cited in Table 2.
66
-------
TABLE 2
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
1.
"STAGS" - Seattle
Xacoma A.erosol
Characterization
Study (Cooper et al.,
1985)
2.
State of Washington
Dept. of Ecology
Monitoring Program
(Beck and Associates,
1984)
3.
"PANORAMAS" - facific
Northwest Eegional
Aerosol Mass
Apportionment Study
(Core et al., 1987)
Site Location
and
Study Period
• Urban sites
in Washington
• 10/82 to
12/82,
2/83 to 3/83,
1/84 to 3/84
• 3 sites in
Washington
• 5/83 to 10/83
• 7 sites in
Washington,
5 sites in
Oregon,
4 sites in
Idaho
• 5/84 to 9/84
4.
"PACS" - £ortland
Aerosol
Characterization
Study (Watson, 1979)
• 6 sites in
Oregon
• 4/77 to 4/78
Study Description
• 24-hour samples for 0-2.5,
2.5-10 and 0-30 Mm size
ranges and analyzed for
mass, elements, ions, and
carbon.
• Chemical mass balance was
applied for SIP
development.
• 24-hour samples for 0-2.5
and 2.5-10 fan size ranges
and analyzed for mass,
elements, ions, and
carbon.
• Chemical mass balance and
factor analysis were
applied.
• 24-hour and 12-hour
samples for 0-1.5 and
0-2.2 (tan size ranges and
analyzed for mass,
elements, ions, and
carbon.
• Concurrent visibility
measurements are also
available for source
apportionment. Chemical
mass balance was applied.
• 24-hour, 8-hour, and 4-
hour samples for 0-2.5 and
0-30 im. size ranges and
analyzed for mass,
elements, ions, and
carbon.
67
• Chemical mass balance was
applied for SIP
development.
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
"MACS" - Bedford
Aerosol
Characterization
Study (DeCesar and
Cooper, 1980)
Site Location
and
Study Period
• 6 sites in
Oregon
• 4/79 to 3/80
"ACHEX" - Aerosol
(Characterization
Experiment (Hidy et
al., 1975)
11 sites in
California
7/72 to 11/72
7/73 to 10/73
Study Description
• 24-hour samples for 0-2,
0-2.5, and 0-30 /ra size
ranges and analyzed for
mass, elements, ions, and
carbon.
• Chemical mass balance was
applied.
• Six 2-hour samples were
collected daily and
analyzed for mass,
elements, ions, and
carbon. Concurrent
meteorological
measurements are also
available.
"SCAQS" - Southern
California &ir
Quality S_tudy
(Blumenthal et al. ,
1987)
"RESOLVE" - Research
on Operations -
Limiting Visual
Extinction (Trijonis
et al., 1987)
• 9 major sites
in Los
Angeles Area
• 11/86, 6/87,
7/87 to 8/87,
11/87 to
12/87
• 8 sites in SE
Desert of
California
• 8/83 to 8/85
• Chemical mass balance was
applied.
• 24-hour and 4-hour samples
for 0-2.5, 0-10, and 0-30
pm size ranges and
analyzed for mass,
elements, ions, carbon,
carbonyls, hydrocarbon,
etc.
24-hour samples for 0-2.5
and 0-10 pm size ranges
and analyzed for mass,
elements, ions, and
carbon.
Chemical mass balance and
other receptor modeling
techniques were applied to
attribute the visibility
degradation in the Mohave
Desert.
68
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
9.
10.
11.
Study
and
Reference
California
Particulate Study
(Flocchini et al.,
1976)
"WOGA" - Western Oil
and £as Association
Aerosol Data Base
(Watson et al.,
1987b)
Site Location
and
Study Period
• 14 sites in
California
• 6/73 to 7/75
• 27 sites in
California
• 1/79 to 12/82
California Air
Resources Board
Dichotomous Sampling
Network (Watson et
al., 1987a)
• 7 sites in
California
• 1/80 to 12/82
1/83 to 6/86
12. South Coast Air
Quality Management
District PM10 Studies
(Gray et al., 1988)
• 7 sites in
Southern
California
• 8/85 to 7/86
8/86 to 12/86
• 3 sites in
Riverside,
California
• 1/88 to 12/88
Study Description
• 24-hour samples for 0.1-
0.65, 0.65-3.6, and 3.6-20
fan size ranges and
analyzed for mass and
elements.
• 24-hour samples for 0-2.5,
2.5-15, 2.5-10, 0-3.5, and
3.5-40 fun. size ranges and
analyzed for mass,
elements, and ions. This
data base includes data
collected from U.S. EPA
and California Air
Resources Board.
• 24-hour samples for 0-2.5
and 2.5-15 fim size ranges
and analyzed for mass and
elements.
• Chemical mass balance and
factor analysis were
applied.
• 24-hour samples for 0-2.0
and 0-10 /tm size ranges
and analyzed for mass,
elements, ions, and
carbon.
• Chemical mass balance was
be applied for SIP
development.
• 24-hour samples will be
collected for 0-2.5 and
0-10 fan around Riverside,
CA, and will be analyzed
for mass, elements, ions,
and carbon.
69
-------
Table 2 (continued)
Ambient Particulate Data Bases
Study
and
Reference
13. "SNAPS" - State of
fievada &ir pollution
Study (Chow et al.,
1988)
Site Location
and
Study Period
• 5 sites in
Nevada
• 1/86 to 3/87
Study Description
• 24-hour and 6-hour samples
for 0-2.5 and 0-10 fas. size
ranges and analyzed for
mass, elements, ions, and
carbon.
14. Southern California
Edison Cottonwood
Cove Study (Bowen et
al., 1986)
• Cottonwood
Cove, NV
• 1/81 to
current
• Chemical mass balance was
applied.
• 24-hour samples for 0-2.5
and 2.5-15 pm size ranges
and analyzed for mass and
elements.
IS. East Helena Source
Apportionment Study
(Houck et al., 1984)
• 11 sites in
Montana
• 1/81 to 4/82
• 24-hour samples for 0-2.5,
2.5-15, and 0-30 pm size
ranges and analyzed for
mass, elements, and
carbon.
16. Kalispell, Montana
Source Apportionment
Study (Olsen, 1987)
• Kalispell,
Montana
• 6/86 to 5/87
17. Arizona Statewide
PM10 SIP Development
(DeNee and Neuroth,
1988)
• 7 sites in
Arizona
• 9/87 to 9/90
• Chemical mass balance was
applied.
• 24-hour samples for 0-2.5
and 2.5-15 /*m size ranges
and analyzed for mass,
elements, and carbon.
• Chemical mass balance was
applied.
• 24-hour and sporadic 4-
hour samples for 0-2.5 and
2.5-10 ^m size ranges and
analyzed for mass,
elements, and carbon.
• Chemical mass balance was
applied for SIP
development.
70
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
18. "VISTTA" - Visibility
Impairment due to
Sulfur transport and
transformation in the
Atmosphere
(Blumenthal et al. ,
1981; Gahill et al. ,
1981 ; Hering et al.,
1981; Macias et al.,
1981a; 1981b)
19. 1978 Denver Winter
Haze Study (Heisler
et al., 1980)
Site Location
and
Study Period
• 2 sites in
Arizona, 2
sites in
Illinois
• 6/79 to 7/29,
12/79, 2/81,
8/81 to 9/81
• 6 sites in
Colorado
• 11/78 to
12/78
Study Description
• 24-hour samples for 0-2.5
and 2.5-15 size ranges and
analyzed for mass,
elements, ions, and
carbon. Concurrent
meteorological
measurements are also
available.
• 4-hour samples for 0-2.5,
2.5-15, and 0-30 pm size
ranges and analyzed for
mass, elements, ions, and
carbon. Concurrent
visibility measurement is
also available.
20. "SCENIC" Denver" -
Study (Cooperative for
Emissions and Impact
Characterization in
Denver (Watson et
al., 1988B)
• 6 sites in
Colorado
• 11/87 to 2/88
Chemical mass balance and
other receptor modeling
techniques were applied.
7 and 17-hour day/night
samples for 0-2.5 fm. size
range and analyzed for
mass, elements, ions, and
carbon. Concurrent
meteorology and visibility
measurements are also
available.
• Chemical mass balance and
other receptor modeling
techniques will be applied
to assess the effect of
visibility degradation due
to emission changes.
71
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
21. "AWVS" - The
Albuquerque Winter
Visibility Study (Zak
et al., 1984)
Site Location
and
Study Period
• 3 sites in
New Mexico
• 12/82 to 2/83
Study Description
• 12-hour daily samples for
0-2.5 and 2.5-10 /un size
ranges and analyzed for
mass, elements, ions, and
carbon.
Multiple linear regression
techniques were applied.
22. North Dakota Study
(Schock et al., 1979)
23. "RAPS" - gegional Air
pollution Study (Loo
et al., 1978; Dzubay,
1980)
• 4 sites in
North Dakota
• 5/87 to 10/87
• 10 sites in
Missouri
• 5/75 to 3/79
24. "SCENES" - Study
Cooperative Electric
Utility, Department
of Defense, Rational
Park Service, and
Environmental
Protection Agency
Study (Mueller et
al., 1986; McDade et
al., 1989)
Regional Scale:
• 9 sites in
California, 3
sites in
Nevada, 3
sites in
Utah, and 4
sites in
Arizona
• 5/85 to
current
• 24-hour daily samples were
collected with stacked
filter units (0-2.0 ^m)
and analyzed for mass and
elements.
• 6-hour and 12-hour samples
for 0-2.5 and 2.5-15 pm
size ranges and analyzed
for mass and elements.
• Chemical mass balance and
other receptor modeling
techniques were applied.
• 8-hour and 24-hour samples
for 0-2.5, 0-10, and 0-15
ion size ranges and
analyzed for mass,
elements, ions, and
carbon. Meteorological
measurements are also
available. A major summer
intensive study was
conducted in summer, 1987,
to develop a regional
profile for the South
Coast Air Basin and to
detect contributions from
this urban area in the
Grand Canyon.
72
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
25.
"WHITEX" - Winter
Haze Intensive Tracer
Experiment (Malm et
al., 1989)
Site Location
and
Study Period
• 2 sites in
Arizona, 2
sites in
Utah, 8
secondary
sites (1 in
Arizona, 7 in
Utah)
• 1/87 to 2/87
26.
27.
"IMPROVE" -
Xnteragency
Monitoring and
Protected Visual
Environments (Joseph
et al., 1987)
"WRAQS" - Western
Regional A.ir Quality
Study (Tombach et
al., 1987)
Regional Scale:
• 36 sites in
United States
(20 IMPROVE
sites and 16
NPS sites)
• 10/87 to 9/90
Regional Scale:
• 10 sites in
California,
Arizona,
Utah,
Colorado,
Montana,
Idaho, New
Mexico, and
Wyoming
• 7/80 to 10/82
Study Description
• 6-hour and 12-hour samples
for 0-2.5 too. size ranges.
Six DRUM samplers were
also operated concurrently
to collect 9 particle size
ranges between 0.07 and
8.5 /on size ranges.
Samples were analyzed for
mass, elements, ions,
carbon, absorption, and
hydrogen.
• Chemical mass balance and
other receptor modeling
techniques will be
applied.
• 24-hour samples are
collected and analyzed for
mass, elements, ions,
carbon, and hydrogen.
• 3-hour, 4-hour, and 8-hour
samples for 0-2.5 and 2.5-
15 fjm size ranges and
analyzed for mass,
elements, ions, and
carbon.
73
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study
and
Reference
28. "WFPS" - Western line
^article Study
(Flocchini et al.,
1981; Cahill et al.,
1981)
29. NFS Partlculate
Monitoring Network
(Cahill et al., 1986)
30. Harvard Air Pollution
Health Effects Study
(Spengler and
Thurston, 1984)
Site Location
and
Study Period
Regional Scale:
• 40 sites in
Montana,
North Dakota,
South Dakota,
Wyoming,
Utah, New
Mexico,
Colorado, and
Arizona
• 10/79 to 5/80
Regional Scale:
• 34 sites in
Washington,
Oregon,
California,
Montana,
Idaho, North
and South
Dakota,
Colorado,
Nevada, Utah,
Arizona, N.
Mexico, Texas
Arkansas,
Tennessee,
and Virginia
• 6/82 to 5/86
Regional Scale:
• 6 sites in
Massachu-
setts,
Tennessee,
Kansas,
Wisconsin,
Ohio, and
Missouri
Study Description
• 72-hour samples for 0-2.5
and 2.5-15 urn size ranges
and analyzed for mass and
elements.
• Factor analysis was
applied to describe the
spatial and temporal
variations.
72-hour samples for 0-2.5
and 2.5*15 fan size ranges
and analyzed for mass,
elements, hydrogen, and
Chemical mass balance was
applied on selected sites.
• 24-hour samples for 0-2.5
and 2.5-15 /an size ranges
and analyzed for mass and
elements.
• Chemical mass balance and
principal component
analysis were applied.
• 2/79 to 7/81
74
-------
TABLE 2 (continued)
AMBIENT PARTICULATE DATA BASES
Study Site Location
and and
Reference Study Period Study Description
31. U.S. EPA Inhalable Urban and • 24-hour samples for 0-2.5
Particulate Matter Regional Scale: and 2.5-15 pm size ranges
Network (Watson et • 73 sites in and analyzed for mass
al., 1981; Rogers and United States elements and ions.
Watson, 1984)
• 5/79 to 6/80 • Chemical mass balance was
applied on selected sites.
75
-------
SECTION 7
REFERENCES
Ahuja, M., J. Paskind, J.E. Houck, and J.C. Chow (1989). "Design of a Study
for the Chemical and Size Characterization of Particulate Matter Emissions
from Selected Sources in California." In Transactions: Receptor Models in
Air Resources Management. J.G. Watson, ed. Air & Waste Management
Association, Pittsburgh, PA, p. 145-158.
Beck, R.W., and Associates (1982). "Washington State Visibility Study."
Prepared for Washington Department of Ecology, Olympia, WA.
Beck, R.W., and Associates (1984). "Washington State Regional Haze/Visibility
Report." Final Report prepared for the Washington Department of Ecology,
Olympia, WA.
Belsley, D.A., E.D. Kuh, and R.E. Welsch (1980). Regression Diagnostics:
Identifying Influential Data and Sources of Col linearity. John Wiley and
Sons, New York, NY.
Benarie, M.M. (1976). "Urban Air Pollution Modeling Without Computers." EPA-
600/4-76-055, U.S. Environmental Protection Agency, Research Triangle Park,
NC.
Blumenthal, D.L., L.W. Richards, E.S. Macias, R.W. Bergstrom, W.E. Wilson, and
P.S. Bhardwaja (1981). "Effects of a Coal-Fired Power Plant on Other Sources
of Southwestern Visibility." Interim Summary of EPA's Project VISTTA. Atmos.
Environ.. 15, 1955.
Blumenthal, D.L., J.G. Watson, P.T. Roberts, and D. Lawson (1987). "Southern
California Air Quality Study: Suggested Program Plan." Report No. 10-95050,
Sonoma Technology, Santa Rosa, CA.
Bowen, J.L., J.G. Watson, and J.C. Chow (1986). "Characterization of the
Southwestern Desert Aerosol." Presented at the 79th Annual Meeting,
Minneapolis, MN. Air Pollution Control Association, Pittsburgh, PA.
Britt, H.I., and R.H. Luecke (1973). "The Estimation of Parameters in
Nonlinear, Implicit Models." Technometrics. 15, 233.
Cahill, T.A., B.H. Kusko, L.L. Ashbaugh, J.B. Barone, and R.A. Eldred (1981).
"Regional and Local Determinations of Particulate Matter and Visibility in the
Southwestern United States During June and July, 1979." Atmos. Environ.. 15,
2011.
76
-------
Cahill, T.A., R.A. Eldred, and P.J. Feeney (1986). "Particulate Monitoring
and Data Analysis for the National Park Service, 1982-1985." Air Quality
Group, University of California, Davis, CA.
Chow, J.C. (1985). "Development of a Composite Modeling Approach to Assess
Air Pollution Source/Receptor Relationships." Doctor of Science Dissertation,
Harvard University, Cambridge, MA.
Chow, J.C., and J.G. Watson (1989). "Summary of Particulate Data Bases for
Receptor Modeling in the United States." In Transactions:Receptor Models in
Air Resources Management. J.G. Watson, ed. Air & Waste Management
Association, Pittsburgh, PA, p. 108-133.
Cooper, J.A., C.A. Frazier, and J.E. Houck (1985). "Seattle-Tacoma Aerosol
Characterization Study (STACS), Volume I, Final Report." NEA, Inc.,
Beaverton, OR.
Cooper, J.A., D.C. Redline, J.R. Sherman, L.M. Valdovinos, W.L. Pollard, L.C.
Scavone, and C. Badgett-West (1987). "PM10 Source Composition Library for the
South Coast Air Basin. Volume I: Source Profile Development Documentation
Final Report." South Coast Air Quality Management District, El Monte, CA.
Core, J.E., N. Maykut, D. Weaver, J. Boylan, and M. Hooper (1987).
"PANORAMAS: Regional Haze in the Pacific Northwest States: A Summary of Major
Findings from the Pacific Northwest Regional Aerosol Mass Apportionment
Study." In Transactions. Visibility Protection: Research and Pol lev Aspects.
P.S. Bhardwaja, ed. Air Pollution Control Association, Pittsburgh, PA.
Core, J.E., J.J. Shah, and J.A. Cooper (1984). "Receptor Model Source
Composition Library." EPA-450/4-85-002, U.S. Environmental Protection Agency,
Research Triangle Park, NC.
Core, J.E. (1989). "Source Profile Development for PM-10 Receptor Modeling."
In Transactions: Receptor Models in Air Resources Management. J.G. Watson,
ed. Air & Waste Management Association, Pittsburgh', PA, p. 134-144.
Currie, L.A., R.W. Gerlach, C.W. Lewis, W.D. Balfour, J.A. Cooper, S.L.
Dattner, R.T. DeCesar, G.E. Gordon, S.L. Heisler, P.K. Hopke, J.J. Shah, G.D.
Thurston, and H.J. Williamson (1984). "Interlaboratory Comparison of Source
Apportionment Procedures: Results for Simulated Data Sets." Atmos. Environ..
18, 1517.
DeCesar, R.T., and J.A. Cooper (1980). "Medford Aerosol Characterization
Study (MACS).11 Final Report to the State of Oregon Department of
Env i ronmental Qua!i ty.
DeCesar, R.T., S.A. Edgerton, M.A.K. Khali!, and R.A. Rasmussen (1985).
"Sensitivity Analysis of Mass Balance Receptor Modeling: Methyl Chloride as
an Indicator of Wood Smoke." Chemosphere. 14, 1495.
77
-------
DeCesar, R.T., S.A. Edgerton, M.A.K. Khalil, and R.A. Rasmussen (1986). "A
Tool for Designing Receptor Model Studies to Apportion Source Impacts with
Specified Precisions." In Transactions. Receptor Methods for Source
Apportionment: Real World Issues and Applications, T.G. Pace, ed. Air
Pollution Control Association, Pittsburgh, PA.
DeNee, P.B., and G. Neuroth (1988). "The White Cloud - Development of a PM10
SIP for a Limestone Processing Facility." Arizona Department of Environmental
Quality, Phoenix, AZ.
Dzubay, T.G. (1980). "Chemical Element Balance Method Applied to Dichotomous
Sampler Data." In Annals of the New York Academy of Sciences. 338, 126.
Dzubay, T.G., R.K. Stevens, VI.D. Balfour, H.J. Williamson, J.A. Cooper, J.E.
Core, R.T. DeCesar, E.R. Crutcher, S.L. Dattner, B.L. Davis, S.L. Heisler,
J.J. Shah, P.K. Hopke, and D.L. Johnson (1984). "Interlaboratory Comparison of
Receptor Model Results for Houston Aerosol." Atmos. Environ.. 18, 1555.
Flocchini, R.G., T.A. Cahill, D.J. Shadoan, S.J. Lang, R.A. Eldred, P.J.
Feeney, and G.W. Wolfe (1976). "Monitoring California's Aerosols by Size and
Elemental Composition." Environ. Sci. Techno!.. 10, 76-82.
Flocchini, R.G., T.A. Cahill, M.L. Pitchford, R.A. Eldred, P.J. Feeney, and
L.L. Ashbaugh (1981). "Characterization of Particles in the Arid West."
Atmos. Environ.. 15, 2017.
Flocchini, R.G., T.A. Cahill, L.L. Ashbaugh, R.A. Eldred, and M. Pitchford
(1981). "Seasonal Behavior of Particulate Matter at Three Rural Utah Sites."
Atmos. Environ.. 15, 315.
Freeman, D.L., N.F. Robinson, J.G. Watson, J.C. Chow, and R.T. Egami (1987).
"Level I PMu Assessment Package User's Manual." Document 8066.2F, prepared
for California Air Resources Board, Sacramento, CA, by Desert Research
Institute, Reno, NV.
Friedlander, S.K. (1973). "Chemical Element Balances and Identification of
Air Pollution Sources." Environ. Sci. Techno!.. 7, 235-240.
Gordon, G.E. (1980). "Receptor Models." Environ. Sci. Technol.. 14, 792.
Gordon, G.E., W.H. Zoller, G.S. Kowalczyk, and S.H. Rheingrover (1981).
"Composition of Source Components Needed for Aerosol Receptor Models." In
Atmospheric Aerosol: Source/Air Quality Relationships. E.S. Macias and P.K.
Hopke, eds. Symposium Series #167, American Chemical Society, Washington, DC.
Gordon, G.E. (1988). "Receptor Models." Environ. Sci. Techno!.. 22, 1132.
Gray, H.A., B. Landry, C.S. Liu, R.C. Henry, J.A. Cooper, and J.R. Sherman
(1988). "Receptor Modeling for PM10 Source Apportionment in the South Coast
Air Basin of California." In Transactions. PM^: Implementation of Standards.
78
-------
C.V. Mathai and D.H. Stonefield, eds. Air Pollution Control Association,
Pittsburgh, PA, p. 399-418.
Heisler, S.L., R.C. Henry, J.G. Watson, and G.M. Hidy (1980). "The 1978
Denver Winter Haze Study." Report P5417-1, prepared for Motor Vehicle
Manufacturers Association of the United States, Inc., by Environmental
Research and Technology, Inc., Westlake Village, CA.
Henry, R.C. (1982). "Stability Analysis of Receptor Models that Use Least
Squares Fitting." In Receptor Models Applied to Contemporary Air Pollution
Problems. S.L. Dattner and P.K. Hopke, eds. Air Pollution Control
Association, Pittsburgh, PA.
Henry, R.C. (1984). "Fundamental Limitations of Factor Analysis Receptor
Models." In Aerosols: Science. Technology and Industrial Applications of
Airborne Particles. B.Y.H. Liu, D.Y.H. Pui, and H.J. Fissan, eds. Elsevier
Press, New York, NY, p. 359.
Hering, S.V., J.L. Bowen, J.G. Wengert, and L.W. Richards (1981).
"Characterization of the Regional Haze in the Southwestern United States."
Atmos. Environ.. 15, 1999-2009.
Hidy, G.M., and S.K. Friedlander (1972). "The Nature of the Los Angeles
Aerosol." 2nd IUAPPA Clean Air Congress, Washington, DC.
Hidy, G.M., B.R. Appel, R.J. Charlson, W.E. Clark, S.K. Friedlander, D.H.
Hutchison, T.B. Smith, J. Suder, J.J. Wesolowski, and K.T. Whitby (1975).
"Summary of the California Aerosol Characterization Experiment." JAPCA. 25,
1106.
Hopke, P.K., and S.L. Dattner, editors (1982). Proceedings. Receptor Models
Applied to Contemporary Pollution Problems. Air Pollution Control
Association, Pittsburgh, PA.
Hopke, P.K. (1985). Receptor Modeling in Environmental Chemistry. John Wiley
& Sons, New York, NY.
Houck, J.E., J.A. Cooper, C.A. Frazier, and R.T. DeCesar (1982). "East Helena
Source Apportionment Study: Particulate Source Apportionment Analysis Using
the Chemical Mass Balance Receptor Model, Volume I: Draft Final Report." NEA
Laboratories, Inc., Beaverton, OR.
Houck, J.E., et al. (1984), "Revised Particulate Source Profiles: East
Helena, Montana." Prepared for ASARCO, Inc.
79
-------
Houck, J.E., J.H. Carlson, C.A. Frazier, and R.A. Cary (1987). "The
Contribution of Residential Wood Combustion to Respirable and Inhalable
Particulate Concentrations in Missoula, Montana." In Proceedings. 1987
EEA/APCA Symposium on Measurement of Toxic and Related Air Polluants. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
Houck, J.E., J.C. Chow, J.G. Watson, C.A. Simons, L.C. Pritchett, J.M. Goulet,
and C.A. Frazier (1989). "Determination of Particle Size Distribution and
Chemical Composition of Particulate Matter from Selected Sources in
California. Volume I." Prepared for California Air Resources Board,
Sacramento, CA, by OMNI Environmental Services, Inc., Beaverton, OR, and
Desert Research Institute, Reno, NV.
Javitz, H.S., J.G. Watson, and N.F. Robinson (1988a). "Performance of the
Chemical Mass Balance Model with Simulated Local-Scale Aerosols." Atmos.
Environ.. 22, 2309-2322.
Javitz, H.S., J.G. Watson, J.P. Guertin, and P.K. Mueller (1988b). "Results
of a Receptor Modeling Feasibility Study." JAPCA. 38, 661.
Joseph, D.B., J. Metsa, W.C. Malm, and M. Pitchford (1987). "Plans for
IMPROVE: A Federal Program to Monitor Visibility in Class I Areas." In
Transactions. Visibility Protection: Research and Policy Aspects. P.S.
Bhardwaja, ed. Air Pollution Control Association, Pittsburgh, PA, p. 113.
Kim, B.M., and R.C. Henry (1989). "Analysis of Multicollinearity Indicators
and Influential Species for Chemical Mass Balance Receptor Model." In
Transactions: Receptor Models in Air Resources Management. J.G. Watson, ed.
Air & Waste Management Association, Pittsburgh, PA, p. 379-390.
Kneip, T.J., M.T. Kleinman, and M. Eisenbud (1972). "Relative Contribution
of Emission Sources to the Total Airborne Particulates in New York City." 3rd
IUAPPA Clean Air Congress, Washington, DC.
Kowalczyk, G.S., C.E. Choquette, and G.E. Gordon (1978). "Chemical Element
Balances and Identification of Air Pollution Sources in Washington, DC."
Atmos. Environ.. 12, 1143.
Lioy, P.J., J.G. Watson, and J.D. Spengler (1980). "APCA Specialty Conference
Workshop on Baseline Data for Inhalable Particulate Matter." JAPCA. 30, 1126.
Loo, B.W., W.R. French, R.C. Gatti, F.S. Goulding, J.M. Jaklevic, and A.C.
Thompson (1978). "Large Scale Measurements of Airborne Particulate Sulfur."
Atmos. Environ.. 12, 759-771.
Macias, E.S., J.O. Zwicker, J.R. Ouimette, S.V. Hering, S.K. Friedlander, T.A.
Cahill, G.A. Kuhlmey, and L.W. Richards (1981a). "Regional Haze Case Studies
in the Southwestern United States--!. Aerosol Chemical Composition." Atmos.
Environ.. 15, 1971-1986.
80
-------
Macias, E.S., J.O. Zwicker, and W.H. White (19815). "Regional Haze Case
Studies in the Southwestern United States--II. Source Contributions." Atmos.
Environ.. 15, 1987.
Malm, W.C., M. Pitchford, and H.K. Iyer (1989). "Design and Implementation of
the Winter Haze Intensive Tracer Experiment - WHITEX." In Transactions:
Receptor Models in Air Resources Management. J.G. Watson, ed. Air & Waste
Management Association, Pittsburgh, PA, p. 432-458.
McDade, C.E., M.R. McGowen, and J.G. Watson (1989). "Identifying Long-Range
Tracers of Los Angeles Emissions." In Transactions: Receptor Models in Air
Resources Management. J.G. Watson, ed. Air & Waste Management Association,
Pittsburgh, PA, p. 407-417.
Mueller, P.K., D.A. Hansen, and J.G. Watson (1986). "The Subregional
Cooperative Electric Utility, Department of Defense, National Park Service,
and Environmental Protection Agency Study (SCENES) on Visibility: An
Overview." Report No. EA-4664-SR, Electric Power Research Institute, Palo
Alto, CA.
Olsen, J.J. (1987). "Final Protocol for a Source Apportionment Study in
Kalispell, Montana." Air Quality Bureau, Montana Department of Health and
Environmental Sciences, Helena, MT.
Pace, T.G., editor (1986). Receptor Methods for Source Apportionment: Real
World Issues and Applications. Air Pollution Control Association, Pittsburgh,
PA.
Pasquill, F. (1974). Atmospheric Diffusion (2nd Ed.)' Ellis Honwood,
England.
Pritchett, L.C., and J.A. Cooper (1985a). "Aerosol Characterization Study of
Anchorage, Alaska: Chemical Analysis and Source Apportionment." Prepared for
Municipality of Anchorage Air Pollution Control Agency, Anchorage, AK, by NEA
Laboratories, Beaverton, OR.
Pritchett, L.C., and J.A. Cooper (1985b). "Aerosol Characterization Study of
Lander and Sheridan, Wyoming: Chemical Analysis and Source Apportionment."
Prepared for State of Wyoming Department of Environmental Quality, Division of
Air Quality, by NEA Laboratories, Beaverton, OR.
Pritchett, L.C., E.A. Miller, C.A. Frazier, and J.A. Cooper (1985c). "Aerosol
Characterization Study of Libby, Montana: Chemical Analysis and Source
Apportionment." Prepared for State of Montana Department of Health and
Environmental Sciences, Air Quality Bureau, Helena, MT, by NEA Laboratories,
Beaverton, OR.
Puget Sound Air Pollution Control Agency (1985). "Seattle-Tacoma Aerosol
Characterization Study (STAGS). Executive Summary." Prepared by NEA, Inc.,
Beaverton, OR.
81
-------
Rogers, C.F., and J.G. Watson (1984). "Potential Causes of Elevated PM10 and
PM15 Concentrations in the Inhalable Particulate Network." Report No. EPA-
450/4-84-016, prepared for U.S. Environmental Protection Agency, Office of Air
Quality and Policy Standards, Research Triangle Park, NC, by Desert Research
Institute, Reno, NV.
Schock, M.R., W.W. Morrison, and G.A. Christiansen (1979). "The Long Term
Effects of Trace Elements Emitted by Energy Conversion of Lignite Coal." Vol.
I. North Dakota State Department of Health Division of Environmental
Research, Bismark, ND.
Seinfeld, J.H. (1975). Air Pollution: Physical and Chemical Fundamentals.
McGraw-Hill, New York, NY.
Shareef, G.S., L.A. Bravo, J.H.E. Stelling, W.B. Kuykendal, and J.D. Mobley
(1989). "Air Emissions Species Data Base." In Transactions: Receptor Models
in Air Resources Management. J.G. Watson, ed. Air & Waste Management
Association, Pittsburgh, PA, p. 73-83.
Sheffield, A.E., and G.E. Gordon (1986). "Variability of Particle Composition
from Ubiquitous Sources: Results from a New Source-Composition Library." In
Transactions. Receptor Methods for Source Apportionment; Real World Issues
and Applications. T.G. Pace, ed. Air Pollution Control Association,
Pittsburgh, PA.
Spengler, J.D., and G.D. Thurston (1984). "Mass and Elemental Composition of
Fine and Coarse Particles in Six U.S. Cities." JAPCA. 33, 1162.
Stevens, R.K., and T.G. Pace (1984). "Review of the Mathematical and
Empirical Receptor Models Workshop (Quail Roost II)." Atmos. Environ.. 18,
1499-1506.
Taback, H.J., A.R. Brienya, J.F. Macho, and N. Brunety (1979). "Fine Particle
Emissions from Stationary and Miscellaneous Sources in the South Coast Air
Basin." Document 5806-783, prepared under Contract AG-19130, for California
Air Resources Board, by KVB, Tustin, CA.
Tombach, I.H., D.W. Allard, R.L. Drake, and R.C. Lewis (1987). "Western
Regional Air Quality Studies; Visibility and Air Quality Measurements:
1981-1982." Report EA-4903, prepared for Electric Power Research Institute,
Palo Alto, CA, by AeroVironment, Inc., Monrovia, CA.
Trijonis, J.C., M. Pitchford, M. McGown, D. Blumenthal, T. Dodson, R.
Flocchini, R. Kelso, W.R. Knuth, E. Macias, J. O'Gara, A. Pitchford, T. Smith,
A. Waggoner, J. Watson, R. Weiss, and W. White (1987). "Preliminary
Extinction Budget Results from the RESOLVE Program." In Transactions.
Visibility Protection: Research and Policy Aspects*. P.S. Bhardwaja, ed. Air
Pollution Control Association, Pittsburgh, PA, p. 872.
Trijonis, J., M. McGown, M. Pitchford, D. Blumenthal, P. Roberts, W. White, E.
Macias, R. Weiss, A. Waggoner, J. Watson, J. Chow, and R. Flocchini (1988).
82
-------
"The RESOLVE Project: Visibility Conditions and Causes of Visibility
Degradation in the Mojave Desert of California. Final Report." Prepared for
Naval Weapons Center/China Lake, Bloomington, CA, by Santa Fe Research
Corporation, Bloomington, MM.
U.S. Environmental Protection Agency (1987a). "Receptor Model Technical
Series, Volume III (Revised). CMB User's Manual (Version 6.0)." EPA-450/4-
83-014R, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC.
U.S. Environmental Protection Agency (1987b). "Protocol for Applying and
Validating the CMB Model." EPA-450/4-86-010, Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
NC.
U.S. Environmental Protection Agency (1987c). "Protocol for Reconciling
Differences Among Receptor and Dispersion Models." EPA-450/4-87-008, Office
of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC.
U.S. Environmental Protection Agency (1988). "Air Emissions Species Manual:
Volume II, Particulate Matter Species Profiles." EPA-450/2-88-003b, Office of
Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC.
Watson, J.G. (1979). "Chemical Element Balance Receptor Model Methodology for
Assessing the Sources of Fine and Total Particulate Matter." Ph.D.
Dissertation, Oregon Graduate Center, Beaverton, OR. University Microfilms
International, Ann Arbor, MI.
Watson, J.G., J.C. Chow, and J.J. Shah (1981). "Analysis of Inhalable and
Fine Particulate Matter Measurements." EPA-450/4-81-035, U.S. Environmental
Protection Agency, Research Triangle Park, NC.
Watson, J.G. (1984). "Overview of Receptor Model Principles." JAPCA. 34, 620-
623.
Watson, J.G., J.A. Cooper, and J.J. Huntzicker (1984). "The Effective
Variance Weighting for Least Squares Calculations Applied to the Mass Balance
Receptor Model." Atmos. Environ.. 18, 1347-1355.
Watson, J.G., and N.F. Robinson (1984). "A Method to Determine Accuracy and
Precision Required of Receptor Model Measurements." In Qua!itv Assurance in
Air Pollution Measurements. Air Pollution Control Association, Pittsburgh,
PA.
83
-------
Watson, J.G., J.C. Chow, D.L. Freeman, R.T. Egami, P. Roberts, and R. Countess
(1987a). "Model and Data Base Description for California's Level I PM10
Assessment Package." Document 8066-002.1F1, Final Report, prepared for the
California Air Resources Board, Sacramento, CA, by Desert Research Institute,
Reno, NV.
Watson, J.G., J.Y. Dea, D.S. Blumenthal, and D.E. Lehrman (1987b). "WOGA
Aerosol Data Base Description." DRI Document 6687.3F2, Final Report, prepared
for the Western Oil and Gas Association, Los Angeles, CA, by Desert Research
Institute, Reno, NV.
Watson, J.G., C.E. McDade, R. Farber, and S. Barr (1987c). "Suggested Program
Plan for a 1987 SCENES Summer Intensive Study, Version 2.1." Document No.
8523-003.2D2, prepared for the SCENES Working Group by Desert Research
Institute, Reno, NV.
Watson, J.G., J.C. Chow, R.T. Egami, C.A. Frazier, A. Goodrich, and C. Ralph
(1988). "PMi0 Source Apportionment in the Truckee Meadows, Nevada, for State
Implementation Plan Development. Volume I: Modeling Methods and Results.
Final Report." DRI Document 8086.2F, prepared as part of The State of Nevada
Air Pollution Study (SNAPS) for the Nevada State Legislature by Desert
Research Institute, Reno, NV.
Watson, J.G., J.C. Chow, L.W. Richards, S.R. Anderson, J.E. Houck, and D.L.
Dietrich (1988b). "The 1987-88 Metro Denver Brown Cloud Air Pollution Study,
Volume III: Data Interpretation." DRI Document No. 8810.1F3, prepared for
the Greater Denver Chamber of Commerce, Denver, CO, by Desert Research
Institute, Reno, NV.
Watson, J.G., editor (1989). Transactions. Receptor Models in Air Resources
Management. Air X Waste Management Association, Pittsburgh, PA.
WDOE and R.W. Beck and Associates (1983). "Washington State Visibility
Study." Report, Olympia, WA.
Williamson, H.J., and D.A. DuBose (1983). "Receptor Model Technical Series,
Volume III: User's Manual for Chemical Mass Balance Model." EPA-450/4-83-014,
U.S. Environmental Protection Agency, Research Triangle Park, NC.
Winchester, J.W., and G.D. Nifong (1971). "Water Pollution in Lake Michigan
by Trace Elements from Aerosol Fallout." Water, Air. Soil Pollut.. 1, 50.
Zak, B.D., W. Einfeld, H.W. Church, G.T. Gay, A.L. Jenson, J. Trijonis, M.D.
Ivey, P.S. Hermann, and C. Tipton (1984). "The Albuquerque Winter Visibility
Study, Volume I: Overview and Data Analysis." National Technical Information
Service, U.S. Department of Commerce, Springfield, VA.
84
-------
APPENDIX A
THEORY OF THE CHEMICAL MASS BALANCE RECEPTOR MODEL
A-l
-------
APPENDIX A
THEORY OF THE CHEMICAL MASS BALANCE RECEPTOR MODEL
A.I INTRODUCTION
Receptor models use the chemical and physical characteristics of gases
and particles measured at source and receptor to both identify the presence of
and to quantify source contributions to the receptor. The particle
characteristics must be such that: 1) they are present in different
proportions in different source emissions; 2) these proportions remain
relatively constant for each source type; and 3) changes in these proportions
between source and receptor are negligible or can be approximated.
Common types of receptor models include: 1) chemical mass balance
(CMB); 2) principal component analysis (PCA, otherwise known as factor
analysis); and 3) multiple linear regression (MLR). Extensive explanations of
each of these models, operating separately and together, are given by Watson
(1984), Chow (1985), Hopke (1985), and Watson et al. (1987a; 1987b). The PCA,
CMB, and MLR have been combined with a dispersion model in a PM,Q assessment
package prepared for the California Air Resources Board (Freeman et al., 1987;
Watson et al., 1987a) which provides interfaces among data bases and modeling
software. The chemical mass balance (CMB) is the fundamental receptor model,
and the derivation of the PCA and MLR models from fundamental physical
principles begins with the CMB.
The chemical mass balance consists of a least squares solution to a set
of linear equations which expresses each receptor concentration of a chemical
species as a linear sum of products of source profile species and source
contributions. The source profile species (i.e., the fractional amount of the
species in the emissions from each source-type) and the receptor
concentrations, with appropriate uncertainty estimates, serve as input data to
the CMB model. The output consists of the amount contributed by each source-
type to each chemical species. The model calculates values for the
A-3
-------
contributions from each source and the uncertainties of those values. Input
data uncertainties are used both to weight the importance of input data values
in the solution and to calculate the uncertainties of the source
contributions.
A.2 DERIVATION, EQUATIONS, AND SOLUTIONS
The concentration measured at a receptor during a sampling period of
length T due to a source j with constant emission rate Ej is
where fT
D,= Jo
S, - D, • E, (A-l)
d [u(t), a (t), xj dt (A-2)
is a dispersion factor depending on wind velocity (u), atmospheric stability
(a), and the location of source j with respect to the receptor (xj. All
parameters in Equation A-2 vary with time, so the instantaneous dispersion
factor, d, must be an integral over time period T (Watson, 1979).
Various forms for d have been proposed (Pasquill, 1974; Seinfeld, 1975;
Benarie, 1976), some including provisions for chemical reactions, removal, and
specialized topography. None are completely adequate to describe the
complicated, random nature of dispersion in the atmosphere. The advantage of
receptor models is that an exact knowledge of D^ is unnecessary.
If a number of sources, J, exists and there is no interaction between
their emissions to cause mass removal, the total mass measured at the
receptor, C, will be a linear sum of the contributions from the individual
sources.
J J
C = E Dj • E, = Z S3 (A-3)
j-l J-l
Similarly, the concentration of elemental component i, C, will be
J
C, = Z F,j • Sj 1 - 1....I (A-4)
A-4
-------
where F,j = the fraction of source contribution S., composed of element
i.
The number of chemical species (I) must be greater than or equal to the
number of sources (J) for a unique solution to these equations.
Solutions to the CMB equations consist of: 1) a tracer solution; 2) a
linear programming solution; 3) an ordinary weighted least squares solution
with or without an intercept; 4) a ridge regression weighted least squares
solution with or without an intercept; and 5) an effective variance least
squares solution with or without an intercept. An estimate of the uncertainty
associated with the source contributions is an integral part of several of
these solution methods.
Weighted linear least squares solutions are preferable to the tracer and
linear programming solutions because: 1) theoretically they yield the most
likely solution to the CMB equations, providing model assumptions are met; 2)
they can make use of all available chemical measurements, not just the so-
called tracer species; 3) they are capable of analytically estimating the
uncertainty of the source 'contributions; and 4) there is, in practice, no such
thing as a "tracer."
CMB software in current use applies the effective variance solution
developed and tested by Watson et al. (1984) because this solution: 1)
provides realistic estimates of the uncertainties of the source contributions
(owing to its incorporation of both source profile and receptor data
uncertainties); and 2) gives greater influence to chemical species with higher
precisions in both the source and receptor measurements are than to species
with lower precisions.
The effective variance solution is derived by minimizing the weighted
sums of the squares of the differences between the measured and calculated
values of C, and F,4 (Britt and Luecke, 1973). The solution algorithm is an
iterative procedure which calculates a new set of S3 based on the S: estimated
from the previous iteration. It is carried out by the following steps
expressed in matrix notation. A superscript k is used to designate the value
of a variable at the kth iteration.
A-5
-------
1. Set initial estimate of the source contributions equal to zero.
=0 j-l,..J (A-5)
k-O
2. Calculate the diagonal components of the effective variance
matrix, V.. All off-diagonal components of this matrix are equal
to zero.
Vktl1 » o|, + 2($y. oju (A-6)
3. Calculate the k+1 value of S:
k+l T k -1 -IT It -1
S - (F (V.) F) F (V.) C (A-7)
4. Test the (k+l)th iteration of the Sj against the kth iteration. If
any one differs by more than 1 percent, then perform the next
iteration. If all differ by less than 1 percent, then terminate
the algorithm.
k+l k k+l
if Sj - Sj /S4 > 0.01 go to step 2
if $r - Sj /s" * 0.01 go to step 5 (A-8)
5. Assign the (k+l)th iteration to S, and osj. All other calculations
are performed with these final values.
T k+l -1 -1 1/2
0SJ - C(F(V. ) F)J J - 1...J (A-9)
where C = (^...C,)7, a column vector with C, as the ith component
S - (Si...Sj)T, a column vector with Sj as the jth component
F » An I x J matrix of F,r the source composition matrix
oct = One standard deviation precision of the C, measurement
oF1J = One standard deviation precision of the F14 measurement
V. = Diagonal matrix of effective variances
A-6
-------
This effective variance solution algorithm is very general, and it
reduces to most of the solutions cited above with the following modifications:
When the oF1j are set equal to zero, the solution reduces to the
ordinary weighted least squares solution.
When the onj are set equal to the same constant value, the
solution reduces to the unweighted least squares solution.
When a column is added to the F,, matrix with all values equal to
1, an intercept term is computed for the variable corresponding to
this column.
When the number of source profiles equals the number of species
(I=J), and if the selected species are present only in a single,
exclusive source profile, the solution reduces to the tracer
solution.
When the matrix
(FT(VJ"V) (A-10)
is re-written as
(FT(VJ~V - »I) (A-ll)
with
-------
Williamson and Dubose (1983) claimed that the ridge regression reduces
coll inearities. Henry (1982) tested the ridge regression solution with respect
to the separation of urban and continental dust and found that the bias
resulted in physically unrealistic negative values for several of the F,r
The ridge regression solution has not been used in the CMB since these tests
were publ ished.
Formulas for the performance measures are:
1 I J
Reduced chi square = Xz = — 2 [(C, - I F^SJVV.,,] (A- 12)
I-J 1=1 j-1
J
Percent Mass = 100(S Sj)/Ct, where Ct denotes the total measured mass
j-l
I
R square - 1 - [ (I-J)X']/[Z C?/V.,,] (A-13)
Modified Pseudo- Inverse Matrix = (F^VJ-'FJ-'F^V.)^ (A-14)
The Singular Value Decomposition of the weighted F matrix is given by
V.»F = UDVT (A- 15}
where U and V are Ixl and JxJ orthogonal matrices, respectively, and where D
is a diagonal matrix with J nonzero and positive elements called the singular
values of the decomposition. The columns of V are called the eigenvectors of
the composition and their components are associated with source types
mentioned earlier in the discussion of the Similarity/Uncertainty Cluster
Display.
The maximum size of the CMB7 work array, which determines how many
source profiles and fitting species can be used by the software, can be
estimated from the following sum:
Work array size (in 4-byte floating point units) =
A-8
-------
11M + 131' + 8J' + 2I'J' + 5IJ + 4J (A-16)
where
M = number of receptor samples
I' = number of species (fitting and floating)
J' = number of source profiles (fitting and floating)
I = maximum number of fitting species
J = maximum number of fitting sources
The work array size must be less than the number which appears after the
introductory banner in CMB7 (see Section 3.2).
A.3 CMB MODEL ASSUMPTIONS
The CMB model assumptions are:
1. Compositions of source emissions are constant over the period of
ambient and source sampling.
2. Chemical species do not react with each other, i.e., they add
linearly.
3. All sources with a potential for significantly contributing to the
receptor have been identified and have had their emissions
characterized.
4. The source compositions are linearly independent of each other.
5. The number of sources or source categories is less than or equal
to the number of chemical species.
6. Measurement uncertainties are random, uncorrelated, and normally
distributed.
A. 4 EFFECTS OF DEVIATIONS FROM CMB MODEL ASSUMPTIONS
Assumptions 1 through 6 for the CMB model are fairly restrictive and
will never be totally complied with in actual practice. Fortunately, the CMB
model can tolerate deviations from these assumptions, though these deviations
increase the stated uncertainties of the source contribution estimates.
The CMB model has been subjected to a number of tests to determine its
abilities to tolerate deviations from model assumptions (Watson, 1979; Gordon
et al., 1981; Henry, 1982; Currie et a!., 1984; Dzubay et al., 1984; Watson
A-9
-------
and Robinson, 1984; DeCesar et al., 1985; Javitz et al.,1988a, 1988b; and
Watson et al., 1987b). These studies all point to the same basic conclusions
regarding deviations from the above-stated assumptions.
With regard to Assumption 1, source compositions, as seen at the
receptor, are known to vary substantially among sources, and even within a
single source over an extended period of time. These variations are both
systematic and random and are caused by three phenomena: 1) transformation
and deposition between the emissions point and the receptor; 2) differences
in fuel type and operating processes between similar sources or the same
source in time; and 3) uncertainties or differences between the source profile
measurement methods. Evaluation studies have generally compared CMB results
from several tests using randomly perturbed input data and from substitutions
of different source profiles for the same source type. The general
conclusions drawn from these tests are as follows:
The error in the estimated source contributions due to biases in
all of the elements of a source profile is in direct proportion to
the magnitude of the biases.
For random errors, the magnitude of the source contribution errors
decreases as the difference between the number of species and
sources increases.
The most recent and systematic tests are those of Javitz et al. (1988b)
which apply to a simple 4-source urban airshed and a complex 10-source urban
airshed. These tests with 17 commonly measured chemical species showed that
primary mobile, geological, coal-fired power plant, and vegetative burning
source-types can be apportioned with uncertainties of approximately 30% when
coefficients of variation in the source profiles are as high as 50%. This
performance was demonstrated even without the presence of unique "tracer"
species such as selenium for coal-fired power plants or soluble potassium for
vegetative burning. In a complex urban airshed, which added residual oil
combustion, marine aerosol, steel production, lead smelting, municipal
incineration, and a continental background aerosol, it was found that the
geological, coal-fired power plant, and background source profiles were
collinear with the measured species. At coefficients of variation in the
source profiles as low as 25%, average absolute errors were on the order of
A-10
-------
60%, 50%, and 130% for the geological, coal-burning, and background sources,
respectively. All other sources were apportioned with average absolute errors
of approximately 30% even when coefficients of variation in the source
profiles reached 50%. Once again, these tests were performed with commonly
measured chemical species, and results would improve with a greater number of
species which are uniquely emitted by the different source types.
With regard to the nonlinear summation of species, Assumption 2, no
studies have been performed to evaluate deviations from this assumption. While
these deviations are generally assumed to be small, conversion of gases to
particles and reactions between particles are not inherently linear processes.
This assumption is especially applicable to the end products of photochemical
reactions and their apportionment to the sources of the precursors. Further
model evaluation is necessary to determine the tolerance of the CMB model to
deviations from this assumption. The current practice is to apportion the
primary material which has not changed between source and receptor. The
remaining quantities of reactive species such as ammonium, nitrate, sulfate,
and elemental carbon are then apportioned to chemical compounds rather than
directly to sources. While this approach is not as satisfying as a direct
apportionment, it at least separates primary from secondary emitters and the
types of compounds apportioned give some insight into the chemical pathways
which formed them. As chemical reaction mechanisms and rates, deposition
velocities, atmospheric equilibrium, and methods to estimate transport and
aging time become better developed, it may be possible to produce
"fractionated" source profiles which will allow this direct attribution of
reactive species to sources. Such apportionment will require measurements of
gaseous as well as particulate species at receptor sites.
A major challenge to the application of the CMB is the identification of
the primary contributing sources for inclusion in the model, Assumption 3.
Watson (1979) systematically increased the number of sources contributing to
his simulated data from four to eight contributors while solving the CMB
equations assuming only four sources. He also included more sources in the
least squares solutions than those which were actually contributors, with the
following results:
A-ll
-------
Underestimating the number of sources had little effect on the
calculated source contributions if the prominent species
contributed by the missing sources were excluded from the
solution.
When the number of sources was underestimated, and when prominent
species of the omitted sources were included in the calculation of
source contributions, the contributions of sources with properties
in common with the omitted sources were overestimated.
When source-types actually present were excluded from the
solution, ratios of calculated to measured concentrations were
often outside of the 0.5 to 2.0 range, and the sum of the source
contributions was much less than the total measured mass. The low
calculated/measured ratios indicated which source compositions
should be included.
When the number of sources was overestimated, the sources not
actually present yielded contributions less than their standard
errors if their source profiles were significantly distinct from
those of other sources. The over-specification of sources
decreased the standard errors of the source contribution
estimates.
Recent research suggests that Assumption 3 should be restated to specify
that source contributions above detection limits should be included in the
CMB. At this time, however, it is not yet possible to determine the
"detection limit" of a source contribution at a receptor since this is a
complicated and unknown function of the other source contributions, the source
composition uncertainties and the uncertainties of the receptor measurements.
Additional model testing is needed to define this "detection limit."
The linear independence of source compositions required by Assumption 4
has become a subject of considerable interest since the publication of Henry's
(1982) singular value decomposition (SVD) analysis. As previously noted, this
analysis provides quantitative measures of collinearity and the sensitivity of
CMB results to specific receptor concentrations. These measures can be
calculated analytically in each application. Henry (1982) also proposed an
optimal linear combination of source contributions that have been determined
to be collinear. 0 »
Other "regression diagnostics" have been summarized by Belsley et al.
(1980) and have been applied to the CMB by DeCesar et al. (1985a, 1985b). Kim
and Henry (1989) show that most of these diagnostics are useless because they
A-12
-------
are based on the assumption of zero uncertainty in the source profiles. Kim
and Henry demonstrate, through the examination of randomly perturbed model
input data, that the values for these diagnostics vary substantially with
typical random changes in the source profiles.
Tests performed on simulated data with obviously collinear source
compositions typically result in positive and negative values for the
collinear source types as well as large standard errors in the collinear
source contribution estimates. Unless the source compositions are nearly
identical, the sum of these large positive and negative values very closely
approximates the sum of the true contributions.
With most commonly measured species (e.g., ions, elements, and carbon)
and source-types (e.g., motor vehicle, geological, residual oil, sea salt,
steel production, wood burning, and various industrial processes), from five
to seven sources are linearly independent of each other in most cases (Javitz
et al., 1988b).
Gordon et al. (1981) found instabilities in the ordinary weighted least
square solutions to the CMB equations when species presumed to be "unique" to
a certain source type were removed from the solution. Using simulated data
with known perturbations ranging from 0 to 20%, Watson (1979) found: llln the
presence of likely uncertainties, sources such as urban dust and continental
background dust cannot be adequately resolved by least squares fitting, even
though their compositions are not identical. Several nearly unique ratios
must exist for good separation."
With regard to Assumption 5, the true number of individual sources
contributing to receptor concentrations is generally much larger than the
number of species that can be measured. It is therefore necessary to group
sources into source-types of similar compositions so that this assumption is
met. For the most commonly measured species, meeting Assumption 4 practically
defines these groupings.
With respect to Assumption 6 (the randomness, normality, and the
uncorrelated nature of measurement uncertainties), there are no results
available from verification or evaluation studies. Every least square*
solution to the CMB equations requires this assumption, as demonstrated by the
derivation of Watson et al. (1984). In reality, very little is known about
A-13
-------
the distribution of errors for the source compositions and the ambient
concentrations. If anything, the distribution probably follows a log-normal
rather than a normal distribution. Ambient concentrations can never be
negative, and a normal distribution allows a substantial proportion of
negative values, while a log-normal distribution allows no negative values.
For small errors (e.g., less than 20%), the actual distribution may not be
important, but for large errors it probably is important. A symmetric
distribution becomes less probable as the coefficient of variation of the
measurement increases. This is one of the most important assumptions of the
solution method that requires testing.
A.5 MODEL INPUT AND OUTPUT DATA
The chemical mass balance modeling procedure requires: 1)
identification of the contributing sources-types; 2) selection of chemical
species to be included; 3) estimation of the fraction of each of the chemical
species which is contained in each source- type (i.e., the source
compositions); 4) estimation of the uncertainty in both ambient concentrations
and source compositions; 5) solution of the chemical mass balance equations,
and 6) validation and reconciliation. Each of these steps requires different
types of data.
Emissions inventories are examined to determine the types of sources
which are most likely to influence a receptor. Principal components analysis
applied to a time series of chemical measurements is also a useful method of
determining the number and types of sources. After these sources have been
identified, profiles acquired from similar sources (Chow and Watson, 1989,
identify most of the available source profiles) are examined to select the
chemical species to be measured. Watson (1979) demonstrates that the more
species measured, the better the precision of the CMB apportionment.
The ambient concentrations of these species, C,, and their fractional
amount in each source-type emission, Fu, are the measured quantities which
serve as CMB model input data. These values require uncertainty estimates,
oc, and 0Mj, which are also input data. Input data uncertainties are used both
to weight the importance of input data values in the solution and to calculate
the uncertainties of the source contributions. The output consists of: 1) the
A-14
-------
source contribution estimates (Sj) of each source-type; 2) the standard errors
of these source contribution estimates; and 3) the amount contributed by each
source-type to each chemical species.
A-15
-------
APPENDIX B
CMB7 ERROR MESSAGES AND CORRECTIVE ACTION
B-l
-------
APPENDIX B
CMB7 ERROR MESSAGES AND CORRECTIVE ACTION
CMB7 will provide error messages which indicate the corrective actions
which must be taken to eliminate the difficulty. Following is a list of the most
common error messages, their most probable cause, and the actions which might
be taken to correct them.
File system error in file CMBOUT Error Code 1034, Status 004
This error occurs when the number of files in the CONFIG.SYS file located
in the root directory is less than 14. Edit this file to contain the line
FILES=14, then re-boot the system.
Error opening (XXXXXX.XXX) What is the name of your
This error occurs when a file name given to the program does not exist or
is indirect, when the filename extender has not been properly specified, or when
the CONFIG.SYS file has specified files less than 14. Verifying and correcting
the input files names in the INXXXXXX.IN7 file will usually solve this problem.
AKT*JEFFIN*AK MATRIX IN SUB-ROUTINE CEB2 NEEDS IMPROVEMENT
This error message is given when a matrix cannot be inverted. This can
happen when the number of fitting sources exceeds the number of fitting species
or when one or more of the profiles have zero values for all of the fitting
species. The problem can be solved by reducing the number of fitting sources
or by increasing the number of fitting species.
Some receptor concentration standard errors are less than or equal to zero.
Weighted regression cannot be done in this case or
FITTING ELEMENT f XX HAS NON-POSITIVE UNCERTAINTY - .00000 PLEASE REPLACE WITH
POSITIVE DETECTION LIMIT PROGRAM TERMINATED Stop- Program terminated.
B-3
-------
CMB7 forces the user to assign uncertainties to his input data. Examine
the input data files to assure that all receptor concentrations have been
assigned non-zero and non-negative precisions. Edit the files to correct them.
No receptor site selected
No samples have been selected by Command 3. Invoke command 3 and examine
the choices. Selected samples are followed by a '*'.
File not found in file Error Code 1032, Status OOOA
The proper filename extension has not been used in the input file names.
Edit the files such that data files have the .TXT extender for CMB7 or the .DAT
extender for CMB 6.0 files,
RECORD NUMBER XX OF SOURCE (or POLLUTANT) NAME FILE XXXXXX>XXX DID NOT MATCH THE
REQUIRED FORMAT (12, 2X, A8, Al) AND WAS IGNORED or RECORD NUMBER XXX OF SOURCE
COMPOSITION FILE XXXXXX.XXX DID NOT MATCH THE REQUIRED FORMAT AND MAS IGNORED
The input data record is improperly formatted. Examine the input data file
and correct the formatting error in a text editor. If the second error prompt
is given, the CMB may run, but it may give erroneous values. The source
composition data should be verified using Command 11.
AN UNIDENTIFIED POLLUTANT CODE I XX WAS FOUND IN FILE XXXXXX>XXX SOURCE -S04
A pollutant code found in the source composition file did not appear in
the pollutant name file using CMB 6.0 formats. The model will ignore this
species.
******** in any output field
The results of the model exceed the limitations of the output formats.
The number is usually meaningless when this appears.
No convergence after 20 iterations - ENTER a carriage return to VIEW RESULTS
This message is given when the effective variance solution has not
converged on a set of source contribution estimates. It is usually associated
with collinear profiles. Change profiles to obtain convergence.
B-4
-------
APPENDIX C
PRINTOUT OF TEST DATA FOR PACS1
C-l
-------
OCD CMB
C>CMB7
MAKE SURE THAT YOUR CAPS LOCK IS ON !
DISK FILE FOR INITIAL INPUT?
IF NOT ENTER CARRIAGE RETURN
IF SO ENTER NAME OF DISK FILE
00 YOU WISH TO RENAME CMBOUT?
IF NOT ENTER A CARRIAGE RETURN.
IF SO ENTER THE FILE NAME.
**************************************************************
U. S. EPA CHEMICAL MASS BALANCE RECEPTOR MODEL
*** IBM-PC CMB7 89338 ***
EPA PROJECT MGRS:
PRINCIPAL AUTHOR:
THOMPSON G. PACE III, PE
QUANG NGUYEN
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
RESEARCH TRIANGLE PARK, NC
(919)-541-5585
DR. JOHN G. WATSON
DESERT RESEARCH INSTITUTE
UNIVERSITY OF NEVADA SYSTEM
(702)-677-3166
CONTRIBUTING AUTHORS:
DR. J.C. CHOW
MR. J.E. CORE
MR. D.A. OUBOSE
MR. QUANG NGUYEN
MR. P.L. HANRAHAN
DR. R.C. HENRY
MR. T.G. PACE
DR. N.F. ROBINSON
DR. H.J. WILLIAMSON
DR. L. WIJNBERG
77540
Initialize size fraction by selecting receptor site
Strike enter to continue
1 PACS1 08/13/77 24 0 COARS
2 PACS1 08/13/77 24 0 FINE
3 PACS2 01/24/78 24 0
4 PACS2 01/24/78 24 0
5 PACS3 08/07/77 24 0
************************
6 PACS3
08/07/77 24 0
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 1
PACS1
PACS1
PACS2
PACS2
PACS3
6 PACS3
08/13/77 24 0
08/13/77 24 0
01/24/78 24 0
01/24/78 24 0
OS/07/77 24 0
08/07/77 24 0
COARS
FINE
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 2
PACS1
PACS1
PACS2
PACS2
PACS3
6 PACS3
08/13/77 24 0
08/13/77 24 0
01/24/78 24 0
01/24/78 24 0
08/07/77 24 0
08/07/77 24 0
COARS
FINE
COARS
FINE
COARS
FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu:
C-3
-------
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMS Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 5
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .97 PERCENT MASS
CHI SQUARE 1.44 OF
DATE: 08/13/77
0 SIZE:
100.6
13
CMB7
COARS
89338
SOURCE
* TYPE
1
3
4
5
8
11
12
13
MAR IN
UDUST
AUTPB
RDOIL
KRAFT
ALPRO
STEEL
FERMN
SCE(UG/M3)
10.8029
9.5985
9.0906
9.7127
12.3265
11.0997
8.1587
9.8720
STD ERR
1.8240
1.2616
1.3961
1.6108
7.9102
2.2441
1.5239
1.6165
TSTAT
5.8131
7.6081
6.5112
6.0296
1.5583
4.9462
5.3538
6.1072
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0*- 8.0/ 80.0+- 8.0/
Strike enter to continue
UNCERTAINTY/SIMILARITY CLUSTERS
160.0*- 11.3
CMB7 89338
SUM OF CLUSTER SOURCES
Strike enter to continue
SPECIES CONCENTRATIONS - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .97 PERCENT MASS
CHI SQUARE 1.44 DF
DATE: 08/13/77
0 SIZE:
100.6
13
CMB7 89338
COARS
SPECIES
Cl
C9
Cll
C12
C13
C14
C16
C17
C19
C20
C22
C23
C24
C25
C26
C28
C29
C30
TOT
F
NA
MG
AL
SI
s
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
I — MEAS CALC RATIO C/M RATIO R/U
T 80.00000*-
. 73400*-
6.33000*-
1.48000*-
4.84000*-
3.27000+-
2.50000*-
4.67000+-
1.12000*-
1.52000*-
.14000*-
. 27700+-
. 00800*-
2.47000+-
5.41000+-
.77900*-
.05000+-
.21400+-
8.00000
. 07300
. 63300
. 14800
.48400
.32700
.25000
.46700
.11200
. 15200
. 01400
.02800
.00100
.24700
.54100
.07800
.00500
.02100
80.46161*-
.50073+-
6.07759*-
1.48676*-
4.40992+-
3.38783*-
2.41256*-
5.19693+-
1.44197*-
1.47195*-
. 13692*-
.34841+-
.24531+-
2.43813+-
4. 11 644+-
.63016+-
.06613+-
.23739+-
6.36458
.22362
.47089
.52389
.55081
.15779
.28216
1.07658
.38620
.10333
. 03204
.07289
.11640
.12229
.30571
.11802
.00683
.03413
1.01+-
.68+-
.96+-
1.00+-
.91+-
1 . 04+-
.97+-
1.11*-
1.29*-
.97*-
.98+-
1.26+-
.13
.31
.12
.37
.15
.11
.15
.26
.37
.12
.25
.29
30.66+-15.05
.99+-
.76+-
.81+-
1.32+-
1.11*-
.11
.09
.17
.19
.19
.0
-1.0
-.3
.0
-.6
.3
-.2
.4
.8
-.3
-.1
.9
2.0
-.1
-2.1
-1.1
1.9
.6
C-4
-------
Strike enter to continue
C35
C82
C201
C202
C203
C204
BR
PB
OC
EC
S04
N03
1
10
1
8
1
52000+-
78000+-
10000+-
68000+-
10000+-
13000+-
.05200
.17800
1.01000
.16800
.81300
.11300
1
8
1
8
51378+-
93077+-
38185+-
34287+-
11920+-
71048+-
.15795
.27347
1.29686
.39741
1.25659
.41362
1
1
99+-
08+-
83+-
80+-
00+-
63+-
.32
.19
.15
.25
.18
.37
-.0
.5
-1.0
-.8
.0
-1.0
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 1
Cl
C9
Cll
4 C12
5 C13
6 C14
7 C16
8 C17.
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 C35
20 C82
Sizes:
TOT
F
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
FINE
T
COARS
T
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu: U
COARS
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 C35
Sizes: F
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
:NE
C-5
-------
20 C82 PB
21 C201 OC
22 C202 EC
23 C203 S04
24 C204 N03
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
FINE COARS
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 C35
20 C82
21 C201
22 C202
23 C203
24 C204
Sizes
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR
PB
OC
EC
S04
N03
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 2
Sizes: FINE COARS
1 1 MARIN * *
2 2 CDUST
3 3 UOUST * *
4 4 AUTPB * *
5 5 ROOIL * *
6 6 VBRN1
7 7 VBRN2
8 8 KRAFT * *
9 9 SULFT
10 10 HOGFU
11 11 ALPRO * *
12 12 STEEL * «
13 13 FERMN * *
14 14 CARBO
15 15 GLASS
16 16 CAR8F
17 17 S04
18 18 N03
C-6
-------
19 19 OC
20 20 EC
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
Sizes: FINE COARS
1 1 MARIN * *
2 2 CDUST
3 3 UDUST *
4 4 AUTPB * *
5 5 RDOIL * *
6 6 VBRN1
7 7 VBRN2
8 8 KRAFT *
9 9 SULFT
10 10 HOGFU
11 11 ALPRO * *
12 12 STEEL * *
13 13 FERMN * *
14 14 CARBO
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu-.
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 5
NO CONVERGENCE AFTER 20 ITERATIONS. ENTER A CARRIAGE RETURN TO VIEW RESULTS
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1 DATE: 08/13/77 CMB7 89338
SAMPLE DURATION 24 START HOUR 0 SIZE: COARS
R SQUARE .90 PERCENT MASS 95.3
CHI SQUARE 4.46 OF 13
SOURCE
* TYPE
1
3
4
5
11
12
13
MARIN
UDUST
AUTPB
RDOIL
ALPRO
STEEL
FERMN
SCE(UG/M3)
12.8176
11.7520
10.9940
12.6953
13.5183
-.3365
14.8274
STD ERR
2.0932
1.2217
1.4994
1.8275
2.4576
.3005
1.5623
TSTAT
6.1235
9.6198
7.3321
6.9467
5.5006
-1.1196
9.4909
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/ 160.0+- 11.3
Strike enter to continue
C-7
-------
UNCERTAINTY/SIMILARITY CLUSTERS
CMB7 89338 SUM OF CLUSTER SOURCES
Strike enter to continue
SPECIES CONCENTRATIONS - SITE: PACS1
SAMPLE DURATION 24 START HOUR
R SQUARE .90 PERCENT MASS
CHI SQUARE 4.46 DF
DATE: 08/13/77 CMB7 89338
0 SIZE: COARS
95.3
13
SPECIES 1 — MEAS CALC RATIO C/M RATIO R/U
Cl
C9
Cll
C12
C13
C14
C16
C17
C19
C20
C22
C23
C24
C25
C26
C28
C29
C30
Strike
C35
C82
C201
C202
C203
C204
TOT
F
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
enter
BR
PB
OC
EC
S04
N03
T
*
*
*
*
*
*
*
*
*
*
*
*
*
*
to
*
*
*
*
*
*
80
6
1
.00000+-
.73400+-
.33000+-
.48000+-
4.84000+-
3
2
4
1
1
2
5
.27000+-
.50000+-
.67000+-
. 12000+-
.52000+-
.14000+-
.27700+-
.00800+-
.47000+-
.41000+-
. 77900+-
.05000+-
.21400+-
8.00000
.07300
.63300
. 14800
.48400
.32700
.25000
.46700
.11200
.15200
.01400
.02800
.00100
.24700
.54100
.07800
.00500
.02100
76
6
1
5
3
2
5
1
1
2
1
.26823+-
.61844+-
.55688+-
. 13935+-
.28780+-
.64568+-
.40091+-
.67512+-
.89774+-
.15141+-
. 14908+-
.44866+-
.01258+-
.55346+-
.53639+-
. 70895+-
.04439+-
. 18537+-
3.64371
.25714
.56405
. 72038
.67072
. 19486
.35376
1.28691
.57918
.11617
.02497
.09527
.00609
. 14687
.24186
. 15383
.00725
.04604
.95+-
.84+-
1.04+-
.77+-
1.09+-
1.11+-
.96+-
1.22+-
1.69+-
.76+-
1.06+-
1.62+-
1.57+-
1.03+-
.28+-
.91+-
.89+-
.87+-
.11
.36
.14
.49
.18
.13
.17
.30
.54
.11
.21
.38
.79
.12
.05
.22
.17
.23
-.4
-.4
.3
-.5
.5
1.0
-.2
.7
1.3
-1.9
.3
1.7
.7
.3
-6.5
-.4
-.6
-.8
continue
1
10
1
8
1
.52000+-
.78000+-
. 10000+-
.68000+-
.10000+-
.13000+-
.05200
.17800
1.01000
. 16800
.81300
.11300
2
8
1
8
1
.61787+-
.26036+-
.11265+-
.43219+-
.38399+-
.03008+-
.19177
.33068
1.39010
.41766
1 . 60400
.49969
1.19+-
1.27+-
.80+-
.85+-
1.04+-
.91+-
.39
.23
.16
.26
.22
.45
.5
1.3
-1.2
-.6
.2
-.2
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 1
1 Cl
2 C9
3 Cll
4 C12
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
Sizes: FINE COARS
TOT T T
F
NA
MG
AL
SI
S
CL
K
CA
*
*
*
*
*
*
TI
C-8
-------
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 CSS
20 C82
V
CR
MN
FE
NI
CU
ZN
8R
PB
Sizes: FINE COARS
TOT T T
F
NA
MG
AL
SI
S
CL
LS
CA
TI
V
CR
MN
FE
NI
CU
ZN
BR *
PB * *
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
1 Cl
2 C9
3 Cll
4 C12
5 C13
6 C14
7 C16
8 C17
9 C19
10 C20
11 C22
12 C23
13 C24
14 C25
15 C26
16 C28
17 C29
18 C30
19 C35
20 C82
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMS Information to Disk
10 Present Computed Averages of CM8 Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 2
Sizes: FINE COARS
1 1 MARIN * *
2 2 CPUST
3 3 UOUST * *
4 4 AUTPB * *
5 5 RDOIL * *
6 S VBRN1
7 7 VBRN2
8 8 KRAFT *
9 9 SULFT
10 10 HOGFU
11 11 ALPRO * *
12 12 STEEL * *
13 13 FERMN * *
14 14 CARBO
C-9
-------
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 8
Sizes: FINE COARS
1 1 MARIN * *
2 2 CDUST
3 3 UOUST * *
4 4 AUTPB * *
5 5 ROOIL * *
6 6 VBRN1
7 7 VBRN2
8 8 KRAFT
9 9 SULFT
10 10 HOGFU
11 11 ALPRO * *
12 12 STEEL *
13 13 FERMN * *
14 14 CARBO
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 9
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 12
C-10
-------
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 4
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 5
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1 DATE: 08/13/77 CMB7 89338
SAMPLE DURATION 24 START HOUR 0 SIZE: FINE
R SQUARE .98 PERCENT MASS 98.7
CHI SQUARE 1.12 OF 13
SOURCE
* TYPE
1
3
4
5
8
11
12
13
MAR IN
UDUST
AUTPB
RDOIL
KRAFT
ALPRO
STEEL
FERMN
SCE(UG/M3)
12.3889
9.5917
10.0835
11.0603
4.6896
10.6023
8.6729
11.8754
STD ERR
2.2457
1.3876
1.4942
1.9239
5.0467
3.5896
1.3771
1.8321
TSTAT
5.5167
6.9127
6.7486
5.7490
.9292
2.9536
6.2979
6.4820
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/ 160.(H- 11.3
Strike enter to continue
UNCERTAINTY/SIMILARITY CLUSTERS CMB7 89338 SUM OF CLUSTER SOURCES
i a
1 5
a
17.078+-
28.139+-
4.241
3.833
Strike enter to continue
SPECIES CONCENTRATIONS - SITE: PACS1 DATE: 08/13/77 CMB7 89338
SAMPLE DURATION 24 START HOUR 0 SIZE: FINE
R SQUARE .98 PERCENT MASS 98.7
CHI SQUARE 1.12 OF 13
C-ll
-------
SPECIES 1— MEAS CALC RATIO C/M RATIO R/U
Cl
C9
Cli
C12
C13
C14
C16
C17
C19
C20
C22
C23
C24
C25
C26
C28
C29
C30
TOT T 80.00000+-
F .88300+-
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
NI
CU
6.93000+-
.43000+-
4.66000+-
3.02000+-
2.95000+-
5.95000+-
1 . 64000+-
1.78000+-
. 08300+-
.37200+-
.31500+-
2.99000+-
4.53000+-
. 76500+-
. 04400+-
ZN .22500+-
8.00000
. 08800
. 69300
. 04300
.46600
.30200
.29500
.59500
.16400
. 17800
.00800
.03700
. 03200
.29900
.45300
.07700
. 00400
.02300
78.96461+-
.67644+-
6.97025+-
1.60951+-
4.02418+-
2.92212+-
3.02466+-
5.69381+-
1 . 73084+-
1.13537+-
. 10088+-
.39757+-
.20976+-
2.82844+-
4.24446+-
.68246+-
.05274+-
.26786+-
4.82449
.24792
. 56446
.62627
.88919
.13329
.31807
1.24836
.46411
.11366
.01630
. 08308
.12151
.14115
.33269
. 13428
.00510
.03966
.99+-
.77+-
1.01+-
3 . 74+-
.86+-
.97+-
1.03+-
.96+-
1 . 06+-
.81+-
1.22+-
1 . 07+-
.67+-
.95+-
.94+-
.89+-
1.20+-
1 . 19+-
.12
.29
.13
1.50
.21
.11
.15
.23
.30
.10
.23
.25
.39
.11
.12
.20
.16
.21
-.1
-.8
.0
1.9
-.6
-.3
.2
-.2
.2
-1..6
1.0
.3
-.8
-.5
-.5
-.5
1.3
.9
Strike enter to continue
CSS
C82
C201
C202
C203
C204
BR
PB
OC
EC
S04
N03
2
7
1
10
41900+-
53000+-
54000+-
42000+-
30000+-
63800+-
.04200
. 25300
.75400
. 14200
1 . 03400
. 06400
2
8
1
9
56133+-
13749+-
50978+-
33S79+-
78819+-
88402+-
1
1
17386
.30300
35632
.34012
47514
.35938
1.34+-
.84+-
1.13+-
.94+-
.95+-
1.39+-
.44
.15
.21
.26
.17
.58
.8
-1.0
.6
-.2
-.3
.7
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 9
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 12
C-12
-------
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 3
1 PACS1 08/13/77 24 0 COARS
2 PACS1 08/13/77 24 0 FINE *
3 PACS2 01/24/78 24 0 COARS
4 PACS2 01/24/78 24 0 FINE
5 PACS3 08/07/77 24 0 COARS
6 PACS3 08/07/77 24 0 FINE
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu: 1
1 PACS1 08/13/77 24 0 COARS *
2 PACS1 08/13/77 24 0 FINE *
3 PACS2 01/24/78 24 0 COARS
4 PACS2 01/24/78 24 0 FINE
5 PACS3 08/07/77 24 0 COARS
6 PACS3 08/07/77 24 0 FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 2
1 PACS1 08/13/77 24 0 COARS *
2 PACS1 08/13/77 24 0 FINE
3 PACS2 01/24/78 24 0 COARS
4 PACS2 01/24/78 24 0 FINE
5 PACS3 08/07/77 24 0 COARS
6 PACS3 08/07/77 24 0 FINE
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 5
SOURCE CONTRIBUTION ESTIMATES - SITE: PACS1 DATE: 08/13/77 CMB7 89338
SAMPLE DURATION 24 START HOUR 0 SIZE: COARS
R SQUARE .97 PERCENT MASS 100.6
CHI SQUARE 1.44 DF 13
C-13
-------
SOURCE
* TYPE
1
3
4
5
a
11
12
13
MAR IN
UOUST
AUTPB
ROOIL
KRAFT
ALPRO
STEEL
FERHN
SCE(UG/M3)
10.6029
9.5985
9.0906
9.7127
12.3265
11.0997
8.1587
9.8720
STD ERR
1.8240
1.2616
1.3961
1.6108
7.9102
2.2441
1.5239
1.6165
TSTAT
5.8131
7.6081
6.5112
6.0296
1.5583
4.9462
5.3538
6.1072
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
80.0+- 8.0/ 80.0+- 8.0/
Strike enter to continue
UNCERTAINTY/SIMILARITY CLUSTERS
160.0+- 11.3
CM87 89338 SUM OF CLUSTER SOURCES
Strike
enter to
continue
SPECIES CONCENTRATIONS - SITE: PACS1
SAMPLE
DURATION
R SQUARE
CHI SQUARE
24
.97
1.44
START
PERCENT
HOUR
MASS
OF
DATE:
0
100.6
13
08/13/77
CMB7 89338
SIZE:
SPECIES 1 — MEAS CALC RATIO
Cl
C9
Cll
C12
C13
C14
C16
C17
C19
C20
C22
C23
C24
C25
C26
C28
C29
C30
Strike
C35
C82
C201
C202
C203
C204
TOT T
F
NA
M6
AL *
SI *
S
CL *
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
enter to
BR
PB
OC
EC
S04
N03
80.
.
6.
1.
4.
3.
2.
4.
1.
1.
.
f
.
2.
5.
.
•
00000+-
73400+-
33000+-
48000+-
84000*-
27000*-
50000*-
67000*-
12000*-
S2000+-
14000+-
27700+-
00800+-
47000+-
41000+-
77900+-
05000+-
21400+-
8.00000
.07300
.63300
.14800
.48400
.32700
.25000
.46700
.11200
.15200
.01400
.02800
.00100
.24700
.54100
.07800
.00500
.02100
80.
e!
1.
4.
3.
2.
5.
1.
1.
2!
4.
.
46161+- 6
50073+-
07759+-
48676+-
40992+-
38783+-
41256+-
19693+- 1
44197+-
47195+-
13692+-
34841*-
24531*-
43813+-
11644+-
63016+-
06613+-
! 23739+-
.36458
.22362
.47089
.52389
.55081
.15779
.28216
.07658
.38620
. 10333
.03204
.07289
.11640
.12229
.30571
.11802
.00683
.03413
1.
1.'
1.'
1.'
1.
.
1.
30.
.
1.
1-
COARS
C/M RATIO
01+-
68+-
96+-
00+-
91+-
04+-
97+-
11*-
29*-
97*-
98+-
26+-
66+-15
99+-
76+-
81+-
32+-
11+-
.13
.31
.12
.37
.15
.11
.15
.26
.37
.12
.25
.29
.05
.11
.09
.17
.19
.19
R/U
.0
-1.0
-.3
.0
-.6
.3
-.2
.4
.8
-.3
-.1
.9
2.0
-.1
-2.1
-1.1
1.9
.6
continue
l!
10.
1.
8.
1.
52000*-
78000+-
10000+-
68000+-
10000+-
13000+-
. 05200
.17800
1.01000
.16800
.81300
.11300
1.
a.
i.
8.
•
51378+-
93077+-
38185+- 1
34287+-
11920+- 1
71048+-
.15795
.27347
.29686
.39741
.25659
.41362
l'.
l'.
99+-
08+-
83+-
80+-
00+-
63+-
.32
.19
.15
.25
.18
.37
-.0
.5
-1.0
-.8
-.0
-1.0
Strike enter to continue
I Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
C-14
-------
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 11
WHAT 00 YOU WANT TO SEE?
ENTER S FOR SOURCE PROFILE OR R FOR RECEPTOR CONCENTRATIONS.
S
DO YOU WANT TO LOOK AT THE WHOLE MATRIX?
IT IS 20 SOURCES BY 24 SPECIES
N
WHICH SOURCE DO YOU WANT?
GIVE SOURCE CODE
4
SOURCE: AUTPB
Cl TOT 1.0000 f- .0000
C9 F .0000 +- .0001
Cll NA .0000 +- .0005
C12 MG .0000 +- .0050
CIS AL .0110 +- .0050
C14 SI .0082 +- .0030
C16 S .0040 +- .0013
C17 CL .0300 +- .0100
C19 K .0007 +- .0003
C20 CA .0125 +- .0050
C22 TI .0000 +- .0010
C23 V .0000 +- .0000
C24 CR .0000 +- .0001
C25 MN .0000 +- .0002
C26 FE .0210 +- .0080
C28 NI .0002 +- .0001
C29 CU .0007 +- .0003
C30 ZN .0035 +- .0013
C35 BR .0500 -*•- .0170
C82 PB .2000 +- .0300
C201 OC . .5000 +- .1000
C202 EC .0380 -c- .0140
C203 S04 .0130 +- .0040
Strike enter to continue
C204 N03 .0091 +- .0030
Strike enter to continue
WHICH SOURCE DO YOU WANT?
GIVE SOURCE CODE
WHAT DO YOU WANT TO SEE?
ENTER S FOR SOURCE PROFILE OR R FOR RECEPTOR CONCENTRATIONS.
OR ARE YOU DONE?
D
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 8
C-15
-------
CALC SPECIES(PER SOURCE)
inUlVlUUHL IVHI1U =
MEAS SPECIES(ALL
SOURCES)
SOURCE NAME
SPECIES
TOT
F
NA
MG
AL
SI
S
CL
K
CA
TI
V
CR
MN
FE
MI
CU
ZN
MAR IN
.133
.000
.670
.344
.000
.000
.140
.908
.133
.098
.000
.000
.000
.000
.000
.000
.000
.000
UDUST
.120
.001
.027
.100
.131
.822
.000
.000
.088
.189
.692
.009
.540
.004
.102
.000
.058
.049
AUTPB
.114
.000
.000
.000
.021
.023
.015
.058
.006
.075
.000
.000
.000
.000
.035
.002
.133
.149
RDOIL
.121
.007
.054
.000
.011
.029
.517
.000
.024
.101
.076
1.206
.571
.002
.053
.668
.146
.182
KRAFT
.154
.000
.103
.000
.007
.005
.163
.077
.044
.029
.000
.000
7.396
.003
.042
.035
.148
.000
ALPRO
.139
.635
.042
.202
.718
.003
.000
.029
.000
.059
.060
.016
.222
.000
.008
.030
.311
.005
STEEL
.102
.000
.016
.358
.011
.125
.064
.032
.067
.333
.117
.018
21.417
.287
.483
.073
.457
.457
FERMN
.123
.039
.048
.000
.013
.030
.067
.009
.925
.084
.032
.009
.518
.691
.038
.000
.071
.268
Strike enter to continue
SPECIES
BR
PB
OC
EC
S04
N03
MAR IN
.041
.000
.000
.000
.131
.000
UOUST
.001
.020
.032
.089
.001
.002
AUTPB
.874
1.021
.450
.206
.015
.073
RDOIL
.002
.006
.067
.179
.577
.056
KRAFT
.013
.000
.193
.132
.180
.000
ALPRO
.026
.000
.000
.106
.023
.000
STEEL
.000
.035
.000
.000
.025
.000
FERMN
.030
.002
.088
.088
.051
.498
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 14
TRANSPOSE OF SENSITIVITY MATRIX
SOURCE NAME
SPECIES
NA
MG
AL
SI
CL
K
CA
TI
V
CR
MN
FE
NI
CU
ZN
MAR IN
1.00
.26
-.02
-.04
.79
.00
.19
-.01
.09
-.18
-.03
-.03
.01
-.27
.09
UDUST
.01
-.04
.02
1.00
-.03
.07
.05
.37
-.03
-.06
-.06
-.12
-.04
-.12
-.11
AUTPB
-.05
.02
.01
-.02
.07
-.03
.06
-.02
.06
-.12
-.05
-.05
.00
-.07
.14
RDOIL
-.05
-.00
-.01
-.02
-.02
-.01
.08
.05
1.00
-.19
-.04
-.10
.79
-.13
.19
KRAFT
.11
-.22
-.18
.03
-.18
.09
-.19
-.04
-.43
.42
-.11
-.00
-.20
.49
-.33
ALPRO
-.01
.09
1.00
-.12
-.00
.00
.01
.01
.01
-.08
-.00
-.12
.01
.25
-.04
STEEL
-.10
.27
-.22
-.19
.05
-.29
.53
-.07
-.07
.28
-.02
1.00
-.01
.40
.49
FERMN
.03
-.11
.09
.04
-.05
.52
-.14
.02
.01
-.14
1.00
-.38
-.03
-.16
.02
C-16
-------
BR
PB
OC
EC
S04
.10
.17
-.43
-.19
-.17
-.02
-.04
.07
.06
.01
.48
1.00
.22
.03
-.16
.03
.09
-.25
.01
.43
-.19
-.46
1.00
.40
.61
.02
-.00
-.11
.04
-.07
-.03
.02
-.39
-.22
-.28
.01
-.05
.14
.10
.07
Strike enter to continue
SPECIES MARIN UDUST AUTPB RDOIL KRAFT ALPRO STEEL FERMN
N03 -.01 .02 .03 .04 -.01 .01 -.16 .27
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Sraph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 13
1 Graph elemental concentrations
2 Graph source profiles
3 Graph source contributions
4 Graph PM10
5 Exit graph menu
Type the line number to select or deselect
0 for page down, U for page up, ENTER for main menu: 1
Hardcopy? Y or N
Y
Plot file name cmbplot.OOl
1
2
3
4
S
Type
0 for
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
the 1
page
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Graph elemental concentrations
Graph source profiles
Graph source contributions
Graph PM10
Exit graph menu
ine number to select or deselect
down, U for
MARIN
CDUST
UDUST
AUTPB
RDOIL
VBRN1
VBRN2
KRAFT
SULFT
HOGFU
ALPRO
STEEL
FERMN
CARBO
GLASS
CARBF
S04
N03
OC
EC
page up, ENTER for main menu: 2
C-17
-------
Toggle selection, Up or Down, Carnage return to exit: 4
1 1 MARIN
2 2 CDUST
3 3 UOUST
4 4 AUTPB
5 5 ROOIL
5 6 VBRN1
7 7 VBRN2
3 8 KRAFT
9 9 SULFT
10 10 riOGFU
11 11 ALPRO
12 12 STEEL
13 13 FERMN
14 14 CARBO
15 15 GLASS
16 16 CARBF
17 17 S04
18 18 N03
19 19 OC
20 20 EC
Toggle selection, Up or Down, Carriage return to exit:
Hardcopy? Y or N
N
1 Graph elemental concentrations
2 Graph source profiles
3 Graph source contributions
4 Graph PM10
5 Exit graph menu
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 3
Hardcopy? Y or N
N
1 Graph elemental concentrations
2 Graph source profiles
3 Graph source contributions
4 Graph PM10
5 Exit graph menu
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu: 5
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4-Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 6
1 PACS1 08/13/77 24 0 COARS *
2 PACS1 08/13/77 24 0 FINE *
3 PACS2 01/24/78 24 0 COARS *
4 PACS2 01/24/78 24 0 FINE *
5 PACS3 08/07/77 24 0 COARS *
C-18
-------
6 PACS3
08/07/77 24 0 FINE
Type the line number to select or deselect
D for page down, U for page up, ENTER for main menu:
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 10
OUTPUT WILL GO TO HARDCOPY.
DO YOU WANT IT DISPLAYED AT YOUR TERMINAL INSTEAD?
Y
FINE COARS TOTAL
CMB SITE DATE SOURCE (UG/M3) (UG/M3) (UG/M3)
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
CMB SITE
PACS1
PACS2
PACS3
01/24/78 MARIN
08/07/77 MARIN
08/07/77 MARIN
AVERAGE
(STD. OEV.)
DATE
SOURCE
01/24/78 UDUST
08/07/77 UDUST
08/07/77 UDUST
AVERAGE
(STD. DEV.)
DATE
SOURCE
01/24/78 AUTPB
08/07/77 AUTPB
08/07/77 AUTPB
AVERAGE
(STD. DEV.)
DATE
SOURCE
01/24/78 ROOIL
08/07/77 RDOIL
08/07/77 RDOIL
AVERAGE
(STD. DEV.)
12.39
-3.42
20.40
9.79
12.12
FINE
(UG/M3)
9.59
1.17
79.18
29.98
42.82
FINE
(UG/M3)
10.08
17.07
23.93
17.03
6.92
FINE
(UG/M3)
11.06
.74
14.01
8.60
6.97
10.60
.31
15.51
8.81
7.76
COARS
(UG/M3)
9.60
54.03
61.89
41.84
28.20
COARS
(UG/M3)
9.09
3.21
35.26
15.85
17.06
COARS
(UG/M3)
9.71
.19
12.45
7.45
6.43
22.99
-3.12
35.91
18.60
19.88
TOTAL
(UG/M3)
19.19
55.20
141.07
71.82
62.62
TOTAL
(UG/M3)
19.17
20.28
59.19
32.88
22.79
TOTAL
(UG/M3)
20.77
.93
26.46
16.06
13.40
C-19
-------
FINE COARS TOTAL
CMB SITE DATE SOURCE (U6/M3) (U6/M3) (UG/M3)
PACS1
PACS2
PACS3
CMB SITE
01/24/78 KRAFT
08/07/77 KRAFT
08/07/77 KRAFT
AVERAGE
(STD. DEV.)
4.69
14.23
-3.53
12.33
.28
9.42
17.02
14.51
5.89
5.13
8.89
7.34
6.29
12.47
5.84
DATE
SOURCE
FINE COARS TOTAL
(UG/M3) (UG/M3) (UG/M3)
PACS1
PACS2
PACS3
01/24/78 ALPRO
08/07/77 ALPRO
08/07/77 ALPRO
AVERAGE
(STD. DEV.)
10.60
-.44
-1.52
2.88
6.71
11.10
4.19
4.46
6.58
3.91
21.70
3.75
2.94
9.46
10.61
FINE COARS TOTAL
CMB SITE DATE SOURCE (UG/M3) (UG/M3) (UG/M3)
PACS1
PACS2
PACS3
01/24/78 STEEL
OB/07/77 STEEL
08/07/77 STEEL
AVERAGE
(STD. DEV.)
8.67
-1.00
-.02
2.55
5.33
8.16
.56
-2.21
2.17
5.37
16.83
-.45
-2.23
4.72
10.53
FINE COARS TOTAL
CMB SITE DATE SOURCE (UG/M3) (UG/M3) (U6/M3)
PACS1 01/24/78 FERMN 11.88 ' 9.87 21.75
PACS2 08/07/77 FERMN .62 -.11 .51
PACS3 08/07/77 FERMN .09 1.41 1.50
AVERAGE 4.20 3.72 7.92
(STD. DEV.) 6.66 5.38 11.99
Strike enter to continue
1 Change Fitting Species
2 Change Fitting Sources
3 Select Samples
4 Advance to Next Sample
5 Calculate Source Contributions
6 Perform Autofit
7 Present Data
8 Present Source Contributions
9 Write CMB Information to Disk
10 Present Computed Averages of CMB Series
11 Present Source Profile or Receptor Concentrations
12 Write Source Contributions to Species to Disk
13 Graph
14 Present Normalized (over species) MPIN Matrix
15 Exit
Type the line number to select: 15
C>
C-20
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
450/4-90-004
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Receptor Model Technical Series, Volume III
(1989 Revision)
CMB7 User's Manual
5. REPORT DATE
January 1990
6. PERFORMING ORGANIZATION CODE
John G. Watson, Norman F. Robinson
Judith C. Chow, Ronald C. Henry, Bongmann Kim,
Quanq T. Nguyen, Edwin L. Meyer. Thompson G. Pace
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND AOORESS
Desert Research Institute, Reno, NV
University Southern California, L.A., CA
U.S. EPA, OAQPS, RTP, NC
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CX-813187-01-1
12. SPONSORING AGENCY NAME AND AOORESS
U. S. EPA
OAQPS, TSD, SRAB (MD-14)
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: Thompson G. Pace and Quang T. Nguyen
16. ABSTRACT
The Chemical Mass Balance (CMB) receptor model uses chemical composition measured in
the source and receptor samples to estimate the relative contributions of different
source categories to ambient particulate concentration.
This manual describes the CMB7 receptor model software. It is designed to allow users
to use the CMB receptor model constructively with a few hour's learning time. Empha-
sizing rapid command of modeling procedures, the manual covers primarily the
mechanical aspects of operating the model. Information on the theoretical basic
principles of CMB receptor modeling is also briefly explained in the appendices of
this manual.
agency
This manual is intended for wide use by State and local air pollution control/person-
nel in developing State Implementation Plans (SIPs) for PM10. The U. S. Environmental
Protection Agency has published a companion document to this manual that should be
consulted for this application. The Protocol for Applying and Validating the CMB
Model, EPA-450/4-87-010, provides guidance on applicability, assumptions and
interpretation of results. This protocol provides a practical strategy for obtaining
valid results.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Receptor models
Chemical Mass Balance
Source apportionment
Least squares
Multiple Linear Regression
Microcomputer Software User's Manual
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report I
Unlimited
21. NO. OF PAGES
125
20. SECURITY CLASS (This page)
Unlimited
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the CPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
TitJe should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g.. Jaic of issue. ilatt- oj
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR IS)
Give name(s) in conventional order (John R. Doe. J. Robert Doc, etc.). List author's affiliation if it differs, from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with. Translation of, Presented at conference of.
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 •words or less) factual summary of the most significant information contained in the report. If (he report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authon/.ed terms that identify the major
concept of the research and are sufficiently specific and precise to be used us index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignments* will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary l-icld/Croup assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19. & 20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (Rev. 4-77) (R.vorse)
------- |