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)

-------