United States Office of Air Quality EPA-450/4-83-014
Environmental Protection Planning and Standards July 1983
Agency Research Triangle Park NC 27711
__
RECEPTOR
MODEL
TECHNICAL
SERIES
VOLUME III
User's Manual For
Chemical Mass
Balance Model
-------
EPA-450/4-83-014
RECEPTOR MODEL TECHNICAL
SERIES
VOLUME III
User's Manual For Chemical
Mass Balance Model
By
Hugh J. Williamson
Dennis A. DuBose
Contract No. 68-02-3513
EPA Project Officer: Warren P. Freas
U.S. EnvfcWvmental
Region 5. library (PL-12J*
7? West Jactatt Boufevwdl 12ft flHI ;7°
Chicago.ll 60604.3690
Prepared For
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise and Radiation
Monitoring And Data Analysis Division
Research Triangle Park, NC 27711
July 1983
-------
11
-------
PREFACE
Receptor Model Technical Series
Volume III
User's Manual for Chemical Mass Balance Model
Recent improvements in air sampling and analytical techniques have
stimulated rapid growth in new techniques of source impact analysis using
receptor models. These models "decode" the chemical fingerprints and
variability of the ambient aerosol to back-calculate source impacts.
Unlike source (dispersion) models that estimate source impacts from
emission rate, meteorology and stack parameters, receptor techniques
estimate source contributions to the total, fine, coarse, or inhalable
particulate mass using data from ambient aerosol measurements.
This document is the third of a series describing how receptor models
can be used by State and local regulatory agencies to identify particulate
source impacts. Volume I (EPA-450/4-81-016a) provides an overview of
Receptor Model Applications, while Volume II (EPA-450/81-016b) focuses
on the Chemical Mass Balance (CMB) technique, model theory and input
requirements.
This report documents an interactive FORTRAN computer program which
performs aerosol source apportionment through weighted least squares with
options to include effective variance and ridge regression features. The
original version of the program, which performed weighted least squares
with the effective variance option, was developed at the Oregon Graduate
Center based on the Doctoral Dissertation of Dr. John Watson. The program
was later modified at Oregon by Dave Torkelson. The current version has
been developed under contract to EPA by Drs. Hugh Williamson and
Dennis DuBose of Radian Corporation. In the latest version, the ridge
regression feature was added, along with various modifications intended
to enhance the ease of use of the program.
The computer program described in this manual is available on magnetic
tape as EPA-450/4-83-014b and can be acquired in either of two ways:
1. Government and nonprofit agencies may obtain a copy of the
program by sending a blank magnetic tape and written request to:
Chief, Technology Development Section
Air Management Technology Branch/MDAD (MD-14)
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attention: Receptor Model Request
2. Others wishing a copy of the program may purchase it from
the National Technical Information Service, Springfield, Virginia 22161,
telephone (703) 557-4650.
m
-------
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 Radian Corporation.
Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
-------
ABSTRACT
This report presents a discussion of aerosol source apportionment.
Chemical mass balance (CMB) equations are used to estimate the contributions
of the important source types to an ambient particulate concentration.
Separate analyses are ordinarily performed for the fine and coarse size
fractions. The basic solution technique discussed is weighted regression
analysis. In this method, the more accurate ambient concentration data are
weighted more heavily to produce a better solution. The analysis technique
can be modified to include the effective variance feature, the ridge regres-
sion feature, or both. In the effective variance feature, the uncertainties
of all input data (ambient and source) are employed in the analysis. Effec-
tive variance calculations produce a weighted least squares solution with re-
fined estimates of the weights. The ridge feature is especially designed
to handle cases in which the particulate emissions from different sources
are chemically similar. Ridge regression potentially allows more resolu-
tion in the source apportionment, in that a more complex set of sources can
often be handled. The report also documents an interactive computer program
which performs source apportionment using the techniques discussed above. A
set of examples of CMB applications is presented to illustrate the various
options of the program.
-------
CONTENTS
Abstract v
Figures viii
Acknowledgements IX
Section 1: Model Overview 1
1.1 Background and Purpose . 1
1.2 Definition of Terms 4
1.3 Overview of Use of Interactive Source
Apportionment Software 8
Section 2: Conceptual Discussion of Source Apportionment 11
2.1 Conceptual Discussion of the Source
Apportionment Problem 11
2.2 Conceptual Discussion of Statistical
Solution Techniques in CMB Analysis 17
Section 3 : Guide to Use of Interactive Source Apportionment Software . 29
3.1 Description of Data Files 30
3.2 Initialization Phase of an Interactive Session .... 34
3.3 Command Phase of an Interactive Session 36
3.4 Detailed Formats'of Data Files 50
3.5 Command Reference Summary 55
3.6 Computer System Considerations 57
Section 4: Example Run 64
4.1 Introduction 64
4.2 Illustration of Commands 64
References 100
Appendices:
A. Source Program Listings 103
B. Selected Flow Diagrams 160
C. Input and Output for Example Run 167
D. Statistical Approach for Source Apportionment 216
-------
FIGURES
Number Page
1-1 Source Apportionment of Ambient Particulate Matter 2
1-2 Conceptual Flowchart of Activities Involved in Performing
Source Apportionments 9
VI1
-------
ACKNOWLEDGEMENTS
The authors acknowledge the valuable contributions of the Environmental
Protection Agency Project Officer, Mr. Warren Freas. Also, the technical
reviews by Messrs. Chuck Lewis and Tom Pace of the EPA and by Mr. Stan Coerr,
Radian Senior Program Manager, are appreciated.
IX
-------
SECTION 1
MODEL OVERVIEW
1.1 Background and Purpose
This report documents an interactive computer program which can be used
to perform an analysis of the sources of ambient aerosol concentrations.
Figure 1-1 summarizes the objectives to be achieved by source apportionment.
The problem to be addressed consists of the presence of objectionable or
hazardous atmospheric particulate matter. Source apportionment provides an
approach for addressing the problem.
"Source apportionment," "source receptor modeling," and "chemical mass
balance (CMB) analysis" are terms which refer to a statistical analysis
through which the major contributors to ambient aerosol pollution are iden-
tified. Specifically, an ambient particulate concentration is apportioned
among its sources. For example, if the concentration was 110 yg/m , the
analysis might indicate that 40 yg/m3 came from source A, 35 yg/m from
source B, etc. The "sources" are source categories such as windblown dust,
automobiles, petroleum refineries, etc. The exact set of sources which
should be included varies geographically. Separate source apportionments
are usually performed for the fine and coarse size fractions. Apportionment
can be made based on the concentrations determined from a single aerosol
sample or based on the average concentrations from a set of samples.
The required chemical inputs consist of (1) the chemical characteristics
of the aerosols from each source type to be considered, and (2) the chemical
characteristics of the ambient aerosols. It is also possible to use the
uncertainties of the input data Ln the analysis. This allows the more
accurate data to be weighted more heavily to produce a better solution.
-------
CD
oo
<_> O
M UJ
o c
W H
BLI CO
en
i j
w o
u as
Bi H
o z
o o
C/j U
1
CQ
CD
cr:
o
co
- Z H-, c/1
d-, j M 2 <
oi < co w 2
UJ Z 3 X
a, < u
a
O
H W U
Z H Z
W < O CO
S i-J 2 W
Z 3 < O
O O OS
I-, hH ^4 3
H H < O
a: os M u~j
O < as
D< a, w
P- H
< u< <
O ~z.
1-1 J O J O
a! ~Z. en H K
< a o as w
as s o- < H
< u s c- E-
= o <
U U b< S
O
OS < F-I 2 < to
til O H U vJ tJ
H i-l 1-1 i-, 3 J
CJ S « CD o 3-,
< W O S i-H Z
05 X CL, 3 H <
< o S as to
EC O d. <
o o o a,
0)
4J
tfl
Oi
4-1
(0
rH
3
o
to
OH
c
•I-l
J2
4-1
c
01
e
o
•H
O
O.
P.
a)
o
3
O
<_)
HH
OS i
CQ
CD
Ct:
Q_
O H
£3
IJ
O
w H
'y, as
D
CO
•H
-------
A statistical model which employs a set of mass balance equations is
used to apportion the particulate matter among its sources. A separate mass
balance equation is employed for each of a set of chemical species. An
additional mass balance equation employs the aerosol concentration for the
size fraction being considered. The inputs (1) and (2), above, contain in-
formation regarding the set of species selected for inclusion in the analysis
and for the aerosol concentration.
The knowledge of which sources are the largest contributors to the
ambient aerosols plays an important role in the formulation of a source-
specific control strategy. Additionally, source apportionment is a valuable
research tool.
The basic solution technique discussed herein is weighted least squares.
In this technique, the more accurate input ambient concentrations are weighted
more heavily to produce a better solution. The following major options re-
garding the solution technique are also available:
(1) Effective variance calculations—a procedure through which the
errors in the complete set of chemical input data are taken into
account in weighting the more accurate data more heavily to produce
a better solution. If effective variances are not used, the uncer-
tainties in only the ambient concentration data are used.
(2) Ridge regression—a feature designed specifically to handle cases
in which the aerosols of two or more sources are chemically simi-
lar. Soil, road dust, and asphalt production is ordinarily such
a set. If conventional statistical techniques are used rather than
ridge regression, it is usually necessary to combine sources with
similar aerosols into groups and apportion the ambient aerosols
among the groups. The separate contributions of the different
sources within a group, however, are not quantified. Ridge re-
gression provides an opportunity to achieve greater resolution
through analysis of a larger set of separate source categories.
-------
Even with ridge regression, however, it is not possible to
separate the influences of sources whose aerosols are chemically
indistinguishable. A discussion of the pros and cons of ridge re-
gression is presented in Section 2.2.
1.2 Definitions of Terms
This subsection presents a set of definitions of technical terms used in
this report. The reader may wish to scan the definitions for an introduction
to the terms and to refer to them while reading the report. Expanded discus-
sions of many of the terms as they apply to source apportionment are presen-
ted in the following sections.
Source receptor analysis—a type of statistical analysis through which an
ambient aerosol concentration is apportioned among its sources. The terms
"source apportionment analysis" and "chemical mass balance (CMB) analysis " are
used in the same sense in this report.
Chemical mass balance equation—an equation in which a mass or concentra-
tion is equated to the sum of its parts. In this context, the ambient aero-
sol concentration is equated to the aerosol concentration due to source one
plus the concentration due to source two, etc. Similar mass balance equa-
tions are used for individual chemical species which make up the ambient
aerosol.
Vector—a one-dimensional array of numbers, e.g., Xi, Xa, ... X. ... X .
Matrix—a two-dimensional array of numbers, e.g., X.., i=l to n, j=l to
m. Notice that a given row or column of a matrix can be thought of as a
vector.
Source signature—a quantification of the chemical characteristics of
the aerosols from a particular source. A .source .signature Lb a vector
-------
whose t value is the fraction of the aerosol from a specific source which
is composed of chemical species i. Suppose, for example, that the soil in
an area consisted of 20% silicon. Then the entry in the source signature
for soil corresponding to silicon would be 0.20.
Source signature matrix—a matrix whose columns are the signatures for
a set of sources.
Standard error—the standard deviation of the error in a quantity. For
example, the standard errors of the ambient concentrations and of the source
signatures are required inputs if a CMS, analysis is performed using weighted
least squares with effective variances.
Error variance—the square of a standard error.
Regression analysis—a statistical method by which a mathematical model
is developed to predict a dependent variable in terms of one or more other
variables. The coefficients in the model are selected so as to minimize the
sum of squares of the differences between the observed and predicted values
in the data set being used. Regression is also referred to as ordinary
least squares.
Regression coefficient—a parameter in a regression model whose value
is estimated from the data.
Weighted least squares—the same as regression analysis, except that
the values of the dependent variable are weighted according to their uncer-
tainties in order to produce a more accurate solution.
Effective variance—the error variance in an equation, taking into
account the uncertainties in both the dependent variable and the predictor
variables. In a refinement of weighted least squares, effective variances
are used in place of the error variances of the dependent variable alone.
-------
The objective is to weight the data according to all input uncertainties in
performing the model development.
Observation—the vector of values of all variables being considered in
an analysis for one case. In regression analysis, for example, an observa-
tion consists of a value of the dependent variable and the corresponding
values of the independent, or predictor, variables.
Sample size—the number of observations used in a statistical analysis.
Iterative solution—an approximate answer to a particular problem
accomplished by computing a succession of estimates of the true answer. If
the procedure is successful, the estimates approach the true answer as the
number of approximations generated increases. Thus, beyond some point, con-
secutive estimates differ very little from each other. As is discussed in
Section 2, an iterative solution is required if effective variances are used,
Iteration—the set of calculations required to produce one of the
estimates in an iterative procedure.
Correlation coefficient (R)—a measure of the strength of the linear
relationship between two variables. If the correlation is one, a perfect
linear relationship exists, and one variable increases as the other in-
creases. If the correlation is minus one, a perfect linear relationship
exists, and one variable increases as the other decreases. If the correla-
tion is zero, no linear relationship exists between the two variables at
all. The square of the correlation coefficient, R2, can be interpreted as
the fraction of the original scatter in one variable that can bo explained
or predicted in terms of the other.
Multiple correlation coefficient (R)—a measure of the strength of the
linear relationship between a dependent variable and two or more other
variables. If the dependent variable can be predicted perfectly from the
-------
other variables, then R is one. If the dependent variable has no linear re-
lationship at all with the other variables, then R is zero. The quantity R/
is the fraction of the scatter in the dependent variable that can be ex-
plained or predicted in terms of the other variables.
L i ne a r c omb in a t i on — a variable X0 is a linear ' combination of another
set of variables Xi , X2 , ... Xn if a set of coefficients A0 , AI , A2 , ... An
exist such that X0 = A0 + AX .
Mul t i c o 1 1 ine ar i t y — a certain condition which affects the accuracy of
the results of a regression analysis. In the simplest case, multicollinear-
ity exists if two predictor variables are highly correlated. In source
receptor modeling, this occurs if two sources have very similar signatures.
In a more general sense, multicollinearity exists if any predictor is nearly
a linear combination of any subset of the other predictors. When strong
multicollinearities are present, conventional regression techniques
typically produce results with large uncertainties.
Variance inflation factor — a measure of the effect of multicollinearity
on a regression coefficient. A variance inflation factor is the increase in
the error variance of a specific regression coefficient due to the effect
of multicollinearity alone. If the variance inflation factor for a par-
ticular coefficient was one, this would indicate that that coefficient was
not affected by multicollinearity at all.
Ridge regression — a type of regression analysis which is specifically
designed to handle cases affected by multicollinearity. In CMB analyses,
ridge regression has enhanced ability to perform apportionments among
sources whose aerosols are chemically similar. However, apportionment among
sources whose aerosols are chemically indistinguishable is not feasible even
with ridge regression. See the discussion in Section 2.2.
-------
Bias—a systematic as opposed to a random error.
1.3 Overview of Use of Interactive Source Apportionment Software
This report documents an interactive computer program which can be used
to perform source apportionment analyses. In each interactive session, one
or more separate source apportionments can be performed at the option of the
user. A brief overview of the procedures for using the program is presented
here. The overview is summarized in Figure 1-2. The procedures are dis-
cussed in detail in Section 3.
Before an interactive session, the required input data must be pre-
pared. The following five data files are required:
List of Source Codes
List of Species Codes
Source Signature Matrix (Fine Particulate Fraction)
Source Signature Matrix (Coarse Particulate Fraction)
Ambient Particulate Concentration Data
The source code file consists of a numeric code and an alphanumeric
code for each source. The species code file requires similar information
for each chemical species included in the analysis. These files are used
only for labeling and indexing purposes. The source signature and ambient
concentration data are used directly in the calculations.
An interactive session has two phases. The first is the initialization
phase, which consists of a series of queries and prompts from the program
that require responses from the user. In this phase, the user selects the
initial set of sources and chemical species to be considered.
In the second phase, the program prompts the user to enter commands. In
response to some commands, the program queries the user for additional
-------
COMMAND PHASE
User Ordinarily
First Selects
Receptor Site,
Date, Size
Fraction
X
/
cn at
•o -3
C cn 4-i cu "O cn cr
co a) cn cj cn cu 4-1 c v-i
g-H-r-l en |H 4J H C O C
E G i-4 IH CO 3 C IH o -ooioi-iuucQ
C_> Cu CU -U «T3CngMHCUCJO>e
cn o a- e cu c IH UH cu H o
CO lHO>O!-lO>OCUUH,-t -H
l-J CU D U -rH -H r-4. -H PU -H O> C -U
cucjoocu.wcoa.4-i Q cn o o.
uccnCiiuoi-H^cu -HO
CM cuQ-uo,c,c:4J4-i
W .C « J-i rH iH Cu 4->33
u u o cu 3 a. e -H ex rH
to cn en 2 _4
o
en
en
u
en
M
H
W
z
0
hH
H
<^
Nl W
t-H en
< x
rH CU
H
M
z
rH
en
4-1
a.
6
O
to
cu
e
CO
^
oo
0
w
cu
^
<4H
•H
O
cu
a
en
0
4J
to
0)
en
3
UH
O
4J
cu
en
rH
CO
•H
4-1
•H
C
r— 1
-a
c
CO
en
CJ
to
D
o
en
CO
Ol
•H
O
CU
Cu
cn
CO
O
•H
e
cu
f-1.
CJ
c
•H
c
O
•H
CO
3
t — i
O
G
IH
O
G
O
•H
4-1
3
rH
o
en
Qt
r«
4_j
Q
PH
Z
CU
tn
0)
rH
•H
&U
CO
4-1
CO
O
CL
U
01
cn
en
cu
T3
o
o
cu
a
to
3
O
en
1
1
en
at
•a
o
c_>
CO
0)
•H
CJ
CU
0.
en
1
1
0)
0
i-i
3
O
en
01
d
•H
fn
1
1
en
cu
to
3
4-1
rd
g)
•H
en
01
o
to
3
O
en
O)
cn
1-1
CO
O
o
1
1
(0
CU
1-1
3
4-1
CO
a
•rH
cn
4-1
C
01
•H
,a
E
<
I
ca
4-1
CO
o
to
O
4->
a.
0)
CJ
cu
an
CO
4-1
c
1
c
o
•H
4-1
'-1
O
a.
a.
o
to
3
O
00
to
O
u-i
l-i
0)
•o
CU
CO
-------
information. Every command and the associated subsequent queries are dis-
cussed in Section 3. Additionally, the HELP command causes the program to
list all the commands and give a brief statement of the purpose of each one.
In the command phase, the user ordinarily first selects the site, date,
and size fraction (fine or coarse) of the ambient particulate concentrations
to be used. It is possible later to alter the initial source or chemical
species list. The CMB comnand causes the computer program to perform a
source apportionment analysis using the current ambient receptor site, date,
and size fraction and the current lists of chemical species and sources.
Various options are available regarding the solution technique.
After a source apportionment is completed, additional problems can be
set up by changing the site, date, size fraction, list of sources, list of
chemical species, or solution technique. It is possible to perform one or a
long series of source apportionment analyses in a single interactive session.
The principal output from a source apportionment consists of:
(1) an estimate of the contribution of each source to the ambient
aerosol concentration, and
(2) the standard errors, i.e., the uncertainties of those estimates.
Additional statistics are also presented to aid the user in interpreting the
results. Outputs from sample problems are presented in Section 4.
In some instances, source apportionment analyses are performed for the
fine and coarse fractions for the same date and site and the same lists of
species and sources. In these cases, it is possible to command the program
to combine the fine and coarse results to produce a source apportionment of
the total particulate concentration.
A post-processor program is also available. The post-processor is used
to compute mean values and other summary statistics for a series of source
apportionments. This feature could be used, for example, if source appor-
tionments had been performed for n set of dates for the same site.
10
-------
SECTION 2
CONCEPTUAL DISCUSSION OF SOURCE APPORTIONMENT
This section presents a conceptual discussion of source apportionment.
The objective here is to provide the user with the technical background
needed to use the source apportionment software. Equations are presented
only to the extent that they are necessary to explain the basic concepts.
The source apportionment problem is discussed in Section 2.1. The solution
techniques are discussed in Section 2.2. The mathematical details are pre-
sented in Appendix D. The inputs to the computer program and the various
options are presented more formally in Section 3.
2.1 Conceptual Discussion of the Source Apportionment Problem
This subsection presents a conceptual discussion of the source apportion-
ment problem. The first step is to define the basic mass balance equation
used. This equation establishes the relationship between the required inputs
and the outputs and serves as a basis for further discussion.
The standard equation used in CMS analysis is as follows:
m
C. = Z F. . S. (1)
1 J-l 1J J
where
C. is the ambient particulate concentration of chemical species i,
F.. is the fraction of the particulate matter emitted from source
type j comprised of species i,
S. is the ambient particulate concentration resulting from source j, and
m is the number of sources.
11
-------
The set of values F.., 1=1, 2, ... , is the source signature for source
j. This vector of values quantifies the chemical composition of the aerosol
from that source. Suppose, for example, that the species Al, Si, K, Ca, Ti ,
and "Ni were being used in the analysis. A small set of elements has been
selected for illustrative purposes.
Also, suppose that the soil in the area being analyzed consisted of
5.6% Al, 20% Si, 1.9% K, 2.6% Ca, 0.32% Ti, and 0% Ni. Then the following
would be the source signature for soil (assume that soil is source number
F . = 0.056
Jo
F_. =0.20
2Jo
F = 0.019
33o
F. . = 0.026
4^o
F_. = 0.0032
Notice that the percents have been converted to decimal fractions. The
chemical species can be ordered in any way desired, as long as the ordering
is the same for all signatures and for the C. values.
The equation given above is a mass balance equation, as can be explained
as follows. The right side of the equation is the sum of a set of terms of
the following form: F. .S . But F. .S. is the ambient concentration of
ij J ij J
species i attributable to source j. This fact is illustrated through the
following numerical example.
Suppose that soil accounted for 40 pg/m3 of ambient particulate matter,
i.e., S. =40. From the source signature, we know that windblown dust is
Jo
20% silicon. Thus, the ambient concentration of silicon due to windblown
dust is
(F )(S. ) = (0.20) (40 yg/m3) = 8 yg/m3 (3)
Jo -1 o
12
-------
Thus, the equation
m
C. = Z F. .S. (4)
i . , 13 j '
J=l
simply states that the ambient concentration of chemical species i equals
the concentration due to source one plus the concentration due to source two,
etc. In other words, the basic equation used is a mass balance equation.
It is possible to add an intercept term S to the standard equation:
o
m
C. = S + Z F..S. (5)
1 ° j=l ^ 3
The term S is not indexed since it has the same value in each mass balance
o
equation. The term S represents a discrepancy between the measured con-
centration C. and the sum of its parts. If all important sources are in-
cluded and there are no severe data errors, the value of S should be small,
less than 0.05 yg/m3 in most applications. Thus, including the term S
simply provides a check. However, it is not required that it be included.
It is also possible to include a mass balance equation corresponding
to the total aerosol concentration C for the size fraction being analyzed.
This equation has the form
m
C = S + Z S. (6)
° j=l
This equation states that the aero_sol concentration is the sum of the inter-
cept and the contributions of the m sources. The intercept S , discussed
above, is uniformly included in or excluded from all mass balance equations,
including this one.
Note that the mass balance equation involving C has exactly the same
form as the others. It can be written
m
CT = So + Z Fi iSi (7)
o ITJ j
13
-------
where
F. . = 1 for all values of j.
iTJ
The quantity ± is the value of the i index for the equation for the aerosol
concentration.
The basic chemical inputs required in CMB analysis include:
(1) the vector of concentrations, C., i=l to n, and C ,
(2) the source signature matrix, F.., i=l to n, j=l to m, and
(3) depending on the type of statistical analysis to be performed,
the standard errors of the C. and F.. values.
i ij
The unknown quantities in the CMB problem are the S. values, j=l to m, and
the intercept S , if it is included.
The sample size is n+1, since there is one observation for each
chemical species and one observation corresponding to C . Generally speak-
ing, increasing the sample size improves the accuracy in the final results
of any statistical analysis, since there, is a greater opportunity for errors
to balance or"average out. While chemical species which serve as tracers for
specific sources are generally the most helpful, species which appear in the
aerosols from several sources should not be excluded. In any case, the sam-
ple size must exceed the number of quantities to be estimated, m+1 if the
intercept is included and m if not. Having a larger sample size also allows
more accurate estimates of the standard errors of the source contributions.
As in any regression analysis, if the number of regression coefficients esti-
mated was almost as large as the sample size, the standard errors could be
artificially low and the R2 value could be artificially high. Further dis-
cussion of the standard errors is presented below and in Appendix D.
There is no reason why all chemical species present in the ambient
aerosols have to be included in the analysis. However, it is important to
include all important source types. If a major source was inadvertently
left out, its aerosol contribution would probably be apportioned among the
14
-------
sources that were included. The sources whose signatures were most similar
to that of the omitted source would be increased by the greatest amount.
Within a given CMB analysis, two sources can be distinguished only if
they have different source signatures. Thus, each of the m "sources" in-
cluded is actually a category of sources which have the same or very similar
source signatures. The categories can include natural sources, such as soil
and, if appropriate, sea salt, and anthropogenic categories, such as auto-
mobiles, petrochemical plants, coal-fired power plants, cement production,
etc.
A single CMB analysis alone can not differentiate among the sources
within a category. For example, the aerosol contribution of cement produc-
tion could be estimated, but the separate contributions of each cement
facility would not be quantified. However, inferences about the effects
of individual sources can be drawn in some instances from wind information.
It is sometimes helpful to perform several separate analyses of cases with
different wind directions.
Whenever possible, it is beneficial to subdivide source categories to
produce more resolution in the final source apportionment. For example,
automobiles with and without catalytic converters can be treated as two
separate sources, since their aerosols are chemically different. Determin-
ing the appropriate set of source categories in a given situation can be a
subtle problem. It is discussed further in the following subsection.
Another issue involves the handling of secondary aerosols. One approach
is to treat secondary aerosols, especially sulfates and nitrates, as
separate "sources." In the source signature for sulfate, for example, all
entries would be zero except the one corresponding to SO., and this entry
would be one; i.e.:
F = 0 , i + L() , and (8)
Jo
Fi j =L
00
15
-------
where
j is the source index corresponding to secondary sulfate, and
i is the chemical species index corresponding to SO,.
If the elemental S concentration were used rather than SO,, then the
4
scheme above would be employed, except i would be the chemical species
index corresponding to S. In the above, it is assumed that not both S and
SO, are included in the Sc?me analysis.
In most applications, ambient concentrations are available for a
series of sampling periods. There are several ways in which these data can
be employed. First, a separate CMS analysis could be performed for each
sampling period. This approach is beneficial if a separate source apportion-
ment for each period is needed. Subsequently, averaging of the CMB results
can be performed. In most such applications, the same source signature
matrix would be used for all periods. Thus, whatever errors are present in
the source signatures are common to all periods. Therefore, the separate
solutions have correlated errors.
The upshot of this is as follows. Suppose a set of values S1. , S2. , ...
H h ^ ^
S. are obtained for S. for H sampling periods. The value S. is the aerosol
J 3 th J
contribution of source j in the h sampling period, 1<_h<_H. The conven-
tional mean value S.
3
o" = i 7 sh
•i H -i
J Hh=l J
is an unbiased estimate of the average aerosol contribution of source j
during the study period. However, the conventional standard error of the
mean
'm I /••_ " { c" _ C \^I * V'U)
m I „„. ,, h=1 Ibj b...
L s Low-biased due to the correlated errors in the individual S. values.
16
-------
The regression analysis for period h produces the estimate S. and the
corresponding standard error s... Another approach is to use error propaga-
tion techniques to compute the standard error of S. in terms of the s
J jh
values, l
-------
additional features are also provided to handle specific aspects of the
problem. These features include the following:
(1) Weighted least squares—a modification of conventional regression
in which the more accurate concentration data are weighted more
heavily to produce a better solution. Weighted least squares is
a standard feature of the program, not an option.
The following two features are options:
(2) Effective variance calculations—a modification of weighted least
squares in which the uncertainties in the source signatures as well
as the uncertainties in the concentration data are taken into
account.
(3) Ridge regression—a technique designed to handle cases with
similar source signatures.
It is possible to select either option (2) or (3), both of them, or neither.
The discussion of the statistical methods presented here is designed to
help the user select the appropriate method for his problem. The mathemati-
cal details are given in Appendix D.
The basic CMB equations
C. = S +1 F..S. ^12^
are linear in the S. values. Thus, a solution could be obtained through
linear regression analysis. In the terms of regression analysis, the C.'s
are the values of the dependent variable, F.. is the i value of the j
predictor variable, S is the intercept, and S. is the j regression co-
efficient.
The solution by conventional least squares is accomplished by choosing
S and the S values to minimize the sum of squares of the differences
° J
between the left and right sides of the mass balance equation above:
18
-------
J (C. - C^2 + (CT - CT)2 (13)
where
m (14)
C. = S + Z F. .S. v '
1 o j=1 ^ J
and
m (15)
r =9 4- T 9 ^J—'/
CT So +.l=1 Sj
A, /\
i.e., C. and C are the estimated or calculated concentrations, and C. and
C are measured concentrations.
In conventional least squares, all data are weighted equally in the
calculations. This is acceptable if the uncertainties of all the data are
about the same. In weighted least squares, the more accurate data are
weighted more heavily to produce a more accurate solution. In weighted
least squares, S and S, values are selected to minimize the following:
- CT)2 (16)
sc sc
i S
where
s is the standard error of C., and
**j. i
i
s is the standard error of C .
T
Weighted least squares employs the uncertainties of the concentration
data only. However, the source signature data also have uncertainties. It
is possible to account for these uncertainties also by using effective
variances (Watson, 1979). The mass balance equation
m
C. = S + Z F S (17)
.]=!•
f.-in be rewrittCMi
19
-------
m
C. - S - Z F..S. = 0 (18)
The effective variance s2 for this equation is the variance of the
entire left side of the final equation, including the effects of the errors
in C. and F.., j=l to m. The exact expression for the effective variance
is given in Appendix D. The main point is that the effective variance is
an estimate of the total uncertainty in an equation, taking into account
the errors in all input data in the equation.
After the effective variances are calculated, S and the S. values are
o j
selected to minimize the following:
Z (C± - C..)2 + (C - CT)2 (19)
i
-2 2
EV± SEVT
Thus, the effective variance approach involves a weighted least squares
analysis but uses refined estimates of the weights.
Appendix D shows that calculation of the effective variances requires
estimates of the S. values, and the S. values are computed in terms of the
effective variances. For this reason, an iterative solution is required it
effective variances are used.
Before discussion of the third feature, ridge regression, a few comments
will be made about the input standard errors. If blank values were entered
for the standard errors of either the ambient concentrations or the source
signatures, zeroes would be stored. A weighted regression analysis cannot
be performed using species whose concentrations have zero standard errors.
If an attempt is made to do such an analysis, the program will print a mes-
sage and abort that particular source apportionment. Nonzero standard
errors are not required, however, for species used only to compare observed
20
-------
and predicted concentrations but not used in the regression analysis (see
Section 3).
The standard errors for the source signatures are used only if the
effective variance feature is selected. Zero values for these standard
errors would not prevent the analysis from being performed. Moreover, a
zero standard error could be correct in some cases. For example, it could
be known that a given species was absent from the aerosols from a particular
source. Then the signature entry for that source and chemical species would
be zero with a standard error of zero.
However, if some zero standard errors were present due to blanks which
were entered only because the standard errors were not known, inaccuracies
could result. The source signature entries with zero standard errors would
be treated as if they were known exactly, while the entries with nonzero
standard errors would be treated as having random errors. Thus, the weight-
ing of the data would not be in accordance with the actual uncertainties.
As a result, the final results would have larger uncertainties than if the
weighting had been performed correctly. The increase in the uncertainties
might not be reflected in the calculated standard errors of the final re-
sults. If all entries for the source signatures had zero standard error
entries, use of the effective variance option would not change the final
answers. The only effect of using effective variances in this case would
be to increase the computer time somewhat.
The third feature which can be added to the solution technique is
called ridge regression analysis. Ridge regression is an advanced statis-
tical technique which is useful when two or more sources have similar sig-
natures. Clearly, if two sources had identical signatures, their contri-
butions to the ambient aerosol could not be separated through CMB analysis.
If the signatures were similar but not identical, separation would be
possible but difficult to accomplish accurately. The problem is referred to
as "multicollinearity." Stated more generally, multicollinearity exists
21
-------
when any source signature is nearly a linear combination of any subset of
the other signatures.
The effect of multicollinearity is to cause large random errors in the
estimates of the S. if conventional techniques are used. If severe multi-
collinearities are present, it is even possible to obtain negative S.
values with large magnitudes.
The variance inflation factor is a convenient measure of the effects of
multicollinearity in a given problem. The variance inflation factor VIF. is
defined as follows:
VIF. = var (S.) (20)
var1 (S.)
where
var (S.) is the error variance of S. if it is calculated through
conventional regression techniques, and
var1 (S.) is the error variance S. would have had in the absence of
multicollinearity.
Thus, VIF. is the increase in the uncertainty (error variance) of S.
due to the effect of multicollinearity alone. If S. was not affected by
multicollinearity at all, VIF would be one. If VIF. was near one, say
between one and 1.5, one could say that the effect oi mu 1 t i <•<> ] 1 j near i tv
on S. was relatively small. Suppose, however, that VIF. was 100. This
would mean that the error variance of S. was increased by a factor of 100,
and the standard error of S. was increased by a factor of 10, by the effect
of multicollinearity alone. It is beneficial to examine the variance infla-
tion factors in each problem to determine whether multicollinearity is a
serious factor.
22
-------
Factor analysis and principal component analysis are additional sta-
tistical methods which can be used to analyze the multicollinearities in a
given data set. While these methods are outside the scope of this report,
they are discussed by Morrison (1967) and other writers on multivariate
statistical methods.
If multicollinearities are present, there are basically two approaches.
One approach involves first identifying the sources whose signatures are
collinear; soil, road dust, and asphalt production ordinarily form such a set.
Then one of the sources can be chosen to represent each set. In the example
given, this would mean that the combined contribution of soil, road dust,
and asphalt production to the ambient aerosol levels would be estimated,
but their individual contributions would be unknown. Some combining of
sources into categories is usually necessary. However, the greater the
number of separate source categories, the greater the resolution is in the
source apportionment.
The second approach is to employ ridge regression, a technique which is
designed to handle cases affected by multicollinearity. In ridge regression,
the goal is to introduce a small bias into the solution in order to achieve
a large reduction in the random error, so that the total error (bias plus
random error) is reduced.
In ridge regression a parameter k is introduced. The exact mathemati-
cal definition of k is given in Appendix D. The important point is that as
k increases, the random error decreases and the bias error increases. The
objective, then, is to pick a value of k near the optimum, where the total,
or bias plus random, errors are minimal. The program computes the solutions
for a set of values of k. At the user's option, the program will select a
value of k according to criteria given in Appendix D. The user also lias the
option of displaying the solutions Tor the complete sot of values of k.
23
-------
The value k=0 is included in the set. When k=0, the ridge solution is
the same as the conventional least squares solution. If this is the best
solution according to a specified set of criteria, it can be selected. Thus,
using the ridge feature does not prevent the conventional least squares
solution from being used, but it allows an additional set of solutions to be
considered.
There is presently not an algorithm which guarantees selection of the
optimal k value. When multicollinearities are present, the regression co-
efficients are typically extremely sensitive to k for small values of k.
However, as k increases, the coefficients stabilize. Hoerl and Kennard
(1970a, b) suggest choosing k from within the stable region, where moderate
changes in k do not affect the coefficients significantly. Thus, slight
deviation from the optimal k would not cause serious errors.
There is no simple formula which tells the user when sources should be
combined into one category and when their separate aerosol contributions
should be estimated. However, the variance inflation factor, discussed
above, is an objective measure which tells the user which sources are
affected by multicollinearity and to what extent.
One rule of thumb which can be used is to determine whether the signa-
tures of two sources differ more than their uncertainties. Theoretically,
a significant difference for only one chemical species would be sufficient
to allow the user to separate the influences of the two sources. Also, the
user should remember that, generally, random errors tend to balance or
"average out" in a statistical analysis. Thus, if two source signatures wrc
. Lfferent by about the uncertainty level for each of several species, it L .-
possible that meaningful separation of the sources could be achieved. The
use of as many species as possible in the analysis allows the maximum oppor-
'unity for the error averaging effect to occur.
24
-------
Another approach is to perform different source apportionment analyses
with different levels of source resolution and then compare the results.
For example, in one analysis, soil and road dust could be combined into a
"crustal" source, and in a second analysis they could be treated separately.
If the estimates of the separate aerosol contributions for soil and road
dust were physically unreasonable or had excessively large standard errors
even when the ridge regression was used, this would indicate that the two
sources should be combined. However, if the estimates of their separate
contributions were physically reasonable and had acceptably small standard
errors, then they should probably be treated as two sources. The standard
errors of the two separate sources in one analysis should be compared to the
standard error for the crustal source in the other.
Regarding the performance of ridge regression, in a survey article on
the subject, Vinod (1978) states the following:
Readers of this review are familiar with multicollinearity problems,
discussed by Farrar and Glauber (1967), Learner (1973), and others.
Hoerl and Kennard's (1970a, b) ridge regression (RR) offers new hope
for avoiding most serious ill-effects of multicollinearity on ordinary
least squares (OLS) regression coefficients. These include "wrong
signs," drastic changes in regression coefficients after minor data
revision or omission of one or two observations, and conflicting con-
clusions from usual significance tests.
In Monte Carlo experiments the true regression coefficients are speci-
fied. The data structure is usually chosen from a real life regression
problem. Random numbers are used to create hundreds of "typical" re-
gression problems. The estimated regression coefficients using RR and
OLS are compared in terms of mean-squared error (MSE), i.e., the average
squared Euclidian distance between the estimate and Lhe parameter. In
the following independent studies by investigators from diverse fields,
the superiority of RR over OLS [ordinary least squares] is almost always
noted; although there is wide disagreement about the "optimum" RR method.
In support of the statements above, Vinod subsequently references sixteen
published Monte Carlo studies and refers to "many unpublished dissertations."
Selection of the optimum RR method is discussed above.
25
-------
The quotation above was chosen from the extensive literature on ridge
regression because (1) it appears in a survey article and summarizes results
from many independent studies and (2) the results to which Vinod refers are
objective in nature, being comparisons of true and estimated regression co-
efficients in Monte Carlo studies. However, no single quotation could repre-
sent all opinions expressed in the literature. In the following quotation,
Draper and Smith (1981) present another view regarding the simulation re-
sults. They use the letter "8" to denote the ridge parameter, referred to
as "k" here.
Ridge Regression Simulations—A Caution
A number of papers containing simulations of regression situations
claim to show that ridge regression estimates are better than least
squares estimates when judged via mean square error. Such claims must
be viewed with caution. Careful study typically reveals that the simu-
lation has been done with effective restrictions on the true parameter
values—precisely the situations where ridge regression is the appro-
priate technique theoretically. The extended inference that ridge re-
gression is "always" better than least squares is, typically, completely
unjustified.
Summary
From this discussion, we can see that use of ridge regression is per-
fectly sensible in circumstances in which it is believed that large
6-values are unrealistic from a practical point of view. However, it
must be realized that choice of 6 is essentially equivalent to an ex-
pression of how big one believes those S's to be. In circumstances
where one cannot accept the idea of restrictions on the B's, ridge re-
gression would be completely inappropriate.
Note that, in many sets of data, where the sizes of the least squares
estimates are acceptable as they are, the ridge trace procedure would
result in a choice of 0 = 0. A value of 0 ^ 0 would be used only when
the least squares results were not regarded as satisfactory.
There is a large and growing literature on the many aspects and gener-
alizations of ridge regression; some selected references are given at
the end of the book.
26
-------
OPINION. Ridge regression is useful and completely appropriate in cir-
cumstances where it is believed that the values of the regression param-
eters are unlikely to be "large" (as interpreted through a^2 or c2
above). In viewing the ridge traces, a subjective judgment is needed
which either (1) effectively specifies one's Bayesian prior beliefs as
to the likely sizes of the parameters, or (2) effectively places a
spherical restriction on the parameter space. The procedure is very
easy to apply and a standard least squares regression program could
easily be adapted by a skilled programmer. Overall, however, we would
advise against the indiscriminate use of ridge regression unless its
limitations are fully appreciated. (The reader should be aware that
many writers disagree with our somewhat pessimistic assessment of
ridge regression.)
Draper and Smith have indicated that ridge regression is "perfectly
sensible in circumstances in which it is believed that large 3-values
[regression coefficients] are unrealistic from a practical point of view."
They also say that "In circumstances where one cannot accept the idea of
restrictions on the B's, ridge regression would be completely inappropriate."
Recall from the discussion above that the regression coefficients are the
estimates of the source contributions.
The question, then, is whether a priori knowledge exists in source
apportionment applications which would allow coefficients with unreasonably
large magnitudes, if present, to be identified. The answer to this question
is yes. Consider the case in which the signatures of two or more sources
are collinear. Typically, if conventional weighted least squares is used,
one of the collinear sources will have a large positive error, and another
source will have a counterbalancing negative error with large magnitude.
The large errors cause the regression coefficients with large magnitudes
referred to by Draper and Smith. The following are two useful criteria for
identifying this situation in source apportionment applications:
(1) In extreme cases, one of the estimated source contributions may
exceed the aerosol concentration for the size fraction being
analyzed. The existence of an upper bound for contributions from
individual sources provides one criterion for identifying
27
-------
excessively large coefficients. To be conservative one could use
the aerosol concentration plus three times its standard error as
the upper bound.
(2) If multicollinearities are present, the counterbalancing negative
errors mentioned above can cause negative estimates of some source
contributions. If negative estimates with large magnitudes were
present, the solution would clearly be physically unreasonable.
Thus, both an upper bound (the aerosol concentration for the size frac-
tion being analyzed) and a lower bound (zero) exist for the regression co-
efficients. If either of these bounds was significantly exceeded, the solu-
tion should be considered unreasonable. Source apportionment, then, is a
context in which criteria exist for recognizing unreasonably large regres-
sion coefficients. Thus, according to Draper and Smith's discussion, ridge
regression should be considered a useful tool in source apportionment appli-
cations. The authors of this report feel that the use of ridge regression
in a given problem can also be justified on other grounds, such as the pres-
ence of large variance inflation factors and sensitivity of both the regres-
sion coefficients and their standard errors to k, especially for small
values of k.
However, Draper and Smith are right that ridge regression should not be
used indiscriminately, as if its application would automatically eliminate
all problems caused by multicollinearity. As is indicated by both articles,
it cannot be guaranteed that ridge regression will produce a better solution
in every application than does conventional least squares. It should also
be remembered, however, that use of the ridge option does not prevent selec-
tion of the solution for k=0 if this solution is the best. The solution for
k=0 is equivalent to the weighted least squares solution.
28
-------
SECTION 3
GUIDE TO USE OF INTERACTIVE SOURCE APPORTIONMENT SOFTWARE
This section discusses the procedures required to use the interactive
source apportionment software. One or more source apportionment analyses
can be performed in a single interactive session.
There are three types of inputs to the CMB program:
Data Files
Commands
Responses to Queries
The data files, which contain the source signatures, ambient concentration
data, and other information, must be set up before an interactive session
is begun. Commands and responses are entered interactively. The user may
enter any of the allowed commands following the prompt "ENTER COMMAND."
Responses to queries are the user replies to specific questions or data
requests from the CMB program.
There are five input data files that are required by the program:
List of Source Codes
List of Species Codes
Source Signature Matrix (Fine Particulate Fraction)
Source Signature Matrix (Coarse Particulate Fraction)
Receptor Concentration Data
A certain degree of incompleteness is possible. For example, if only fine
fraction data are available for source signatures, an empty data file is
acceptable for the coarse fraction.
An interactive session with the CMB program consists of two phases: an
initialization phase and a command phase. The program controls the
29
-------
initialization phase by prompting the user for the initial sources and
species. The user may add or delete sources and species for either size
fraction in the working phase. The user controls the command phase by issu-
ing commands. The command may be performed immediately or the CMB program
may prompt the user for additional information before proceeding. The user
must spell out the command in full. However, responses to prompts from the
computer may be abbreviated by the first letter of the response unless a
numerical value or alphanumeric code (site code) is requested. For example,
a "Y" has the same effect as "YES." In fact, the program reads only the first
letter of nonnumeric, non-code responses. Thus, "SOURCE SIGNATURE" can be
abbreviated "SOURCE" or "S."
In the following subsections, discussions are presented on the following:
(1) required data files,
(2) initialization phase of the interactive session,
(3) command phase of the interactive session,
(4) formats of required data sets,
(5) command reference summary, and
(6) computer system considerations
3.1. Description of Data Files
The CMB program requires 11 files for input or output. These include
data input files, hardcopy printout files, interactive communications, and
temporary data storage files. These files are accessed by the CMB program
according to the following FORTRAN Input/Output unit numbers:
Unit_# Purpose I/O
2 Source Names/Codes Input
3 Species Names/Codes Input
5 Interactive Read Input
6 Interactive Write Output
7 Fine Fraction Signature Input
8 Coarse Fraction Signature Input
9 Hardcopy Print (Summary) Output
10 Temporary Storage Input/Output
11 Hardcopy Print (General) Output
12 User-Followup Data Output
13 Receptor Concentrations Input
30
-------
The user's terminal serves as the interactive read and write units.
The user's commands and replies are read and the program's queries and out-
put are written. The program output includes the fitted CMB coefficients
and predicted concentrations. Sections 3.2 and 3.3 describe the commands,
queries and replies. Section 4 presents an annotated interactive session.
3.1.1 Input Data Files
There are five input data files that must be available to the program
before an interactive session can begin. The purpose and content of each
data file are described below. The detailed record formats are given in
Section 3.4, and examples are given in Appendix C.
The source and species names are for the convenience of the user. The
source and species codes must be consistent across all data sets. These
codes are used for labeling the data sets only. During the interactive
session the program will assign source and species numbers that may not
agree with the source and species codes. The source and species numbers,
not the codes, must be used in the interactive session. The source and
species numbers can be listed at the beginning of the interactive session.
The source and species codes and numbers, the source signature matrix, and
the current receptor concentrations can be listed with the PMATRIX command
during an interactive session.
The following are descriptions of the required input files:
Source and Source Code List—Eight-character source names and two-
digit source codes are required. The source signature matrix must be
coded consistently with this list. The two-digit codes may range from "01"
to "99" in any order. Gaps are allowed in the code range. Sources not
used in the current source signature matrix can be present. The maximum
number of sources which can be included in the input data files is 35.
31
-------
The maximum number of sources which can be included in a given CMB analysis,
however, is 16.
Species and Species Code List—Eight-character species names and two-
digit species codes are required. The source signature matrix and the re-
ceptor concentrations must be coded consistently with this list. The same
comments made for source codes apply here. The maximum number of species
which can be included in the input data files is 35. The maximum number of
species which can be used in a given CMB analysis is 21.
Fine Fraction Source Signature Matrix—The fraction of the aerosols
from each source accounted for by each species is required. If the effec-
tive variance method is to be used, the standard error of each item is also
required. Each item and its standard error are entered on a separate line
(record or card) along with the identifying source and species codes.
Coarse Fraction Source Signature Matrix—The same comments as for the
fine fraction apply here.
If coarse fraction data are not available, a dummy data set may be
substituted.
Receptor Concentration Data—The receptor concentration and its stan-
dard error are required for each species for both fine and coarse fractions.
Alternatively, the fine fraction and total aerosol may be substituted and
the program will calculate the coarse fraction data. Each species is coded
on a separate line (card or record).
If only fine fraction data are available the coarse fraction may be
coded as zero.
32
-------
3.1.2 Output: Data Files
The CMB program writes to three permanent files: the general hardcopy,
the summary hardcopy, and the user followup data file. The two hardcopy
files are for printed results (132-character line). The postprocessor data
file makes certain input and output data available for further processing.
Examples are given in Appendix C.
General Hardcopy—The general hardcopy printout includes hardcopy
versions of results displayed in the interactive sessions. Some additional
calculated values are presented that are not included in the terminal dis-
plays. The primary output to the general hardcopy is the display of CMB
coefficients with measured and calculated concentrations. Appropriate
standard errors, percentages, and ratios are included. This output results
from a WRITE command.
Certain other commands can direct results to the general hardcopy. The
PMATRIX command prints input matrices. The PSOLN command prints ridge re-
gression solutions. The PCOMP command prints averages of source contribu-
tions.
Summary Hardcopy—The summary hardcopy printout tabulates the CMB
coefficients for fine, coarse and total fractions. Standard errors and
percent composition are included. This report is printed by the WRITE
command only when both fine and coarse fractions are analyzed for the same
site and date.
User Followup Data File—The user followup data may be used for
further analysis with user-written software. The calculated concentration
of each species contributed by each source is written to this file. The
program writes these values to the file by the WRITE command only when both
fine and coarse fractions are analyzed for the same site and date (as with
33
-------
the summary hardcopy). The measured concentrations are also placed on the
file. The detailed format of the data file is given in Section 3.4.
3.2 Initialization Phase of an Interactive Session
The initialization phase consists of a series of queries and prompts
from the program that require responses from the user. Most of the required
responses are to identify the sources and species. Each query is given
below followed by a discussion of the appropriate user response.
WOULD YOU LIKE TO LOOK AT THE SOURCE AND SPECIES LISTS?
Unless the user is familiar with the program and data, he should reply
"YES." If the user responds "YES" the program lists the sources and species
by number and name at the terminal. In other parts of the session the user
will be asked what sources or species he wants to add, delete, or display.
The user must respond with the source or species number according to these
lists. Generally, the program will display the source or species number
along with the name whenever it appears.
PLEASE INPUT INITIAL FITTING INFORMATION
INITIAL SOURCE 1: XX
At this point the user tells the program which sources to include in
the initial problem set. The user should key in the desired source numbers,
one line at a time in response to the program prompts, "INITIAL SOURCE
1: XX," "INITIAL SOURCE 2:XX," and so forth. When all of the desired
sources have been entered, the user should reply with a blank or null
line (carriage return with no data).
If the user wants all the available sources in the initial problem set,
he should enter "-1" and the program will automatically include them up to
its limit, which is 16. The user can also use this feature to save time
34
-------
when almost all sources are wanted. The DS (Delete Source) command can be
used later to remove unwanted sources.
PLEASE INPUT SPECIES
INITIAL SPECIES 1:XX
At this point the user tells the program which sources to include in
the initial problem set. The above discussion for entering initial
sources applies also to species. The DE (Delete Species) command can be
used later to remove unwanted species.
ARE INITIAL SOURCES AND SPECIES CORRECT?
If the user responds "YES," then the session enters the command phase.
The program responds with "ENTER COMMAND" (see below). If the user
responds "NO," then he is prompted with the following query:
USE COMMANDS AE, DE, AS, DS FOR CHANGES.
OR, DO YOU WANT A FRESH START?
Both fine and coarse size fractions will start with the initially
chosen sources and species. The user can use the add and delete commands
to change the list of sources and species for either size fraction. Later
the user can restore either size fraction to the initially chosen sources
and species. A "YES" response to this query returns the session to "PLEASE
INPUT INITIAL FITTING INFORMATION" to re-enter the initial choice of
sources and species. A reply of "NO" gets the following prompt:
35
-------
THE HELP COMMAND LISTS COMMANDS
SIZE IS FINE
ENTER COMMAND
The session has entered the command phase. This phase is discussed
in the next section. The size fraction is currently set to "fine."
3.3 Command Phase of an Interactive Session
The command phase of the CMB program is indicated by the prompt
"ENTER COMMAND." Each command is invoked by entering its name in the key-
board. For some commands this is all that is necessary. Other commands
will prompt the user for additional information.
During the command phase many parameters remain set at previous values
until changed by the user. This feature relieves the user from the repeti-
tive entries that would otherwise be necessary. The parameters that hold
their values and the commands that change them are the following:
Parameters
List of Sources
List of Species
Receptor Concentrations
Size Fraction
Background
Commands to Change
AS, DS, INITIAL
AE, DE, INITIAL
SELECT, AUTOFIT
SIZE, SELECT, AUTOFIT
BACKOUT, BACKIN
The command "PINFO" will display the current status ot these parameters.
3.3.1 HELP
The HELP command lists all the commands with a brief statement of the
purpose of each one.
36
-------
3.3.2 SELECT
The SELECT command is used to select the receptor data for the
current problem. SELECT also automatically invokes commands CMS and PDATA.
Under normal use, SELECT will be the first command of the interactive
session.
SELECT will prompt the user to identify the receptor data as follows:
ENTER DESIRED CMB SITE CODE: XXXXXXXXXXXX
The user should respond with the site code or name. If the sites
available are not known, enter a "?" or "Q" and then give blank responses
to the next three queries. The program will list the available receptors
and dates for which data are available. Subsequently, the user can use
the SELECT command again and choose a site and date from those listed.
ENTER YEAR: YY
Key in the last two digits of the desired year.
ENTER DATE: MMDD
Key in the month and day numbers of the desired sampling date. For
example, June 26 is 0626.
INPUT DESIRED SIZE FRACTION: (FINE OR COARSE)
Key in FINE for the fine fraction or COARSE for the coarse fraction.
The CMB program will now search the receptor concentration data set
for the desired data. As it does so it will display the cases it encounters
37
-------
in the search. At the far right will be the letters FC for Fine/Coarse
data type or FT for Fine/Total data type. The portion not present will be
calculated.
The program then begins the CMB command followed by PDATA.
3.3.3 CMB
The CMB command performs the analysis. It also prompts the user for
the various method options. The method options are:
Weighted Ridge Regression/Weighted Least Squares
Effective Variance Method/No Effective Variances
Fitted Model Intercept/Zero Intercept
These options may be exercised in any combination.
DO YOU WANT RIDGE SOLUTIONS?
Reply "YES" for weighted ridge regression or "NO" for weighted least
squares. (Reply "SAME" for the same method options chosen in the previous
use of the CMB command. The program will then proceed directly to the
analysis without further queries.) If the reply is "YES" then the following
two queries will be presented:
CMB WILL SELECT THE BEST RIDGE SOLUTION
DO YOU WANT TO SEE A SUMMARY OF THE OTHERS?
A "YES" response will cause the program to display the following
items for each ridge k-value: R2, calculated total aerosol (sum of
coefficients), negative coefficient of greatest magnitude, the number of
negative coefficients, and the number of iterations. If the effective
variance option is not used, the number of iterations is always one.
38
-------
IN THE SELECTION OF THE BEST RIDGE SOLUTION, WHAT WEIGHT DO YOU WANT?
(DEFAULT =1.0)
In the selection of the best ridge solution two criteria are considered:
closeness of the calculated aerosol total to the measured aerosol total
and the smallness in magnitude of any negative coefficients. These two
criteria are weighted in choosing the best solution. A weight of 1.0
gives an "even" weight to the two criteria. A weight less than 1.0 gives
more weight to the closeness of the total aerosol calculated and measured
values. A weight greater than 1.0 gives more weight to the smallness in
magnitude of any negative coefficients. The mathematical details for the
selection of the ridge solution are discussed in Appendix D.
Key a null line (press return) for the default weight 1.0. Otherwise,
enter a number above zero but less than ten million. The decimal point
must be keyed.
DO YOU WANT TO USE EFFECTIVE VARIANCES?
A "YES" reply will invoke the effective variance method. Standard
errors must be present on the source signature matrix. Reply "NO" if the
effective variance method is not desired.
DO YOU WANT TO INCLUDE THE INTERCEPT IN THE MODEL?
A "NO" reply sets the intercept to zero. A "YES" reply allows the
program to fit the intercept to the data. A fitted intercept value greatly
different from zero is an indicator of difficulties. This may result from
an important source being omitted or data errors.
The CMB program now performs the analysis. If the effective variance
and ridge options are in effect, the following prompt may occur, particularly
for the first ridge k-value (zero or least squares):
-------
XXX ITERATIONS SO FAR FOR RIDGE K = X.XXX WITHOUT CONVERGENCE. HOW MANY
MORE DO YOU WANT TO TRY?
If the effective variance option is in effect but the ridge option is not,
the following prompt may occur:
XXX ITERATIONS SO FAR WITHOUT CONVERGENCE. HOW MANY MORE DO YOU WANT TO TRY?
Reply with the number of extra iterations. Generally, just a few
more are required (less than 10). If no more iterations are wanted key a
zero, and the solution will be the current estimate.
The program now enters the PDATA command automatically.
3.3.4 PDATA
The PDATA command displays the problem solution at the user's
terminal. The output consists of the solution coefficients and standard
errors. The least squares solution is automatically printed along with
variance inflation factors. This means that the same solution will be given
twice if'the ridge regression option is not used.
Optionally, the actual and calculated receptor concentrations will
be displayed:
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
A null line (press the transmit or enter key) will result in the
optional receptor concentration display. Key in a C if this information
is to be suppressed.
The output from PDATA can be sent to hardcopy with the WRITE command.
40
-------
3.3.5 WRITE
The write command writes the solution and other information to storage
files for later use or display. The files to which data are written are:
General Hardcopy
Summary Hardcopy
User Followup Data, and
Temporary Storage for use by the PCOMP command.
The hardcopy files may be printed at the user's terminal or on the system
printer after the end of the CMB interactive session. Section 3.6 dis-
cusses access to these files.
The report written to the general hardcopy is an expanded version of
the terminal display given by the PDATA command. The report written to the
summary hardcopy is a summary of these reports for fine, coarse and total
sizes. The items written to the user followup data file are contributions
from each source to the calculated concentration of each species.
The WRITE command writes the current CMB analyses results to the
general hardcopy file. This is confirmed at the user's terminal by the
message "WRITTEN." The WRITE command writes to the summary hardcopy, the
user followup data, and the temporary storage only under the following
conditions: both fine and coarse CMB analyses are performed in succession
and each is followed by a WRITE command. The same site and date must be
used for both size fractions. This occurrence is confirmed at the user's
terminal by two successive "WRITTEN" messages. A CMB analysis display
for the total of the fine and coarse fractions is also written to the
general hardcopy at this time.
Section 3.1.2 discusses the contents of the three output files.
Section 3.4 gives the format of the user followup data file. Appendix C
gives examples of the output to these files. A flow diagram illustrating
the WRITE command is given in Appendix B.
41
-------
3.3.6 SIZE
The SIZE command changes the size fraction from the current setting
to the other. (The two alternatives are fine and coarse.) The command
then displays the new current size, fraction.
3.3.7 AE
The AE command is used to add species to the current working list.
The program prompts the user for species to add. The user responses are
checked for validity. After the AE command is made, the user is prompted
with the following:
SIZE IS XXXXXX
INPUT CODE OF ADDED SPECIES
The user should reply with the species number. The program will then
issue the prompt again and another species may be added. When all species
desired are added, give a blank or null line (press transmit or enter) .
If the entry is invalid or unintelligible the program will print a
message and issue a new prompt.
If the current size fraction displayed is not the desired one, then
the user should terminate the command (enter a blank line), give the SIZE
command, and re-enter the AE command.
3.3.8 DE
The DE command is used to delete species from the current working list,
The program prompts the user for species to delete. The user responses
are checked for validity. After the DE command, the user is prompted with
the following:
42
-------
SIZE IS XXXXXX
INPUT CODE OF DELETED SPECIES
The user should reply with the species number. The program will then
issue the prompt again and another species may be deleted. When all species
desired are deleted, give a blank or null line.
If the entry is invalid or unintelligible the program will print a
message and issue a new prompt.
If the current size fraction displayed is not the desired one, then
the user should terminate the command (enter a blank line), give the SIZE
command, and re-enter the DE command.
3.3.9 AS
The AS command is used to add sources to the current working list.
This command functions similarly to AE.
3.3.10 DS
The DS command is used to delete sources from the current working list.
This command functions similarly to DE.
3.3.11 INITIAL
The INITIAL command restores the current size fraction source and
species lists to the initially chosen lists. The initial lists were chosen
in the initialization phase.
3.3.12 PSOLN
The PSOLN command displays all 31 ridge regression solutions. The
display may be directed to the user's terminal or to hardcopy.
-------
The output from this command can be extensive. The program first
warns the user by giving the size of the entire solution set. Then the
program prompts the user as follows:
ARE YOU SURE YOU WANT TO LOOK AT THESE NUMBERS?
Reply "YES" if still serious. A "NO" will get the prompt ENTER
COMMAND.
DO YOU WANT THE SOLUTIONS PRINTED ON THE HARDCOPY? (INSTEAD OF YOUR
TERMINAL)
This option allows the user to get the 31 solutions and examine them
later. A "YES" response sends the solutions to hardcopy. A "NO" response
will display them at the user's terminal.
CAN YOUR TERMINAL DISPLAY A 132 CHARACTER LINE?
Some printer terminals can print a wide 132-character line. Other
printer terminals and some CRT's display a 72-character line. On these
terminals a 132-character line will "wrap around" and be difficult to
read. The program will format the line to either a 132 or a 72-character
line, depending on the user response.
3.3.13 PMATRIX
The PMATRIX command displays the source signature matrix, a single
source signature, the receptor concentrations, or the source and species
codes. The displays may be directed to the user's terminal or to hardcopy.
MATRICES CAN GO TO HARDCOPY
DO YOU WANT THEM DISPLAYED AT YOUR TERMINAL INSTEAD?
44
-------
A "NO" will send all output to hardcopy. Nothing will be displayed
at the user's terminal. A "YES" will result in all output being displayed
at the terminal. The command can be re-entered later to redirect output.
WHAT DO YOU WANT TO SEE?
SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS, OR CODES?
OR ARE YOU DONE?
A reply of DONE or a null line will bring the prompt ENTER COMMAND.
Otherwise, reply "SOURCE" or "RECEPTOR" or "CODES." The reply "CODES"
will display the source and species numbers and input data set codes. The
reply "RECEPTOR" will display the receptor concentrations for both fine
and coarse size fractions. The reply "SOURCE" will result in the following
prompt:
WHAT SIZE FRACTION? (FINE OR COARSE)
Reply FINE for fine fraction source signature data or COARSE for
coarse fraction source signature data.
DO YOU WANT TO LOOK AT THE WHOLE MATRIX?
The program will warn the user of its dimensions. A "NO" response will
allow the user to select individual source signatures for display, A "YES"
response will result in the following query (if output is directed to
user's terminal):
CAN YOUR TERMINAL DISPLAY A 132 CHARACTER LINE?
The program will adjust the format to fit either a 72 or 132-character
line. See PSOLN command for details.
45
-------
The source signature matrix is displayed so that the standard error
is directly below each matrix item.
If the option to display individual source signatures was selected,
the program prompts,
WHICH SOURCE DO YOU WANT?
Key in the source number. The program will display the indicated
source data. The prompt will then be reissued. Another source may be
displayed, or enter a blank or null line to be returned to the prompt WHAT
DO YOU WANT TO SEE?
3.3.14 PINFO
The PINFO command displays the current status of the problem-solving
process. This is more useful for CRT users than those with printer
terminals.
The command displays the sources and species currently assigned. It
also gives the receptor identifier, data, size fraction, and background
information. An example of the display appears in Section 4.
3.3.15 AUTOFIT
The AUTOFIT command is an alternative to the SELECT command. The
SELECT command selects a single receptor site, date, and size fraction for
CMB analysis. In contrast, the AUTOFIT command automatically sequences
through a series of sites, dates, and size fractions. The AUTOFIT command
is equivalent to a series of SELECT and WRITE commands. However, user
responses are minimized by eliminating repetitive keying of commands and
responses.
46
-------
Before using the AUTOFIT command, the user should be familiar with the
SELECT command. Like SELECT, the AUTOFIT command invokes the CMB and
PDATA commands. It also invokes the WRITE command.
Like the SELECT command, the AUTOFIT command queries the user for site,
year, and date. This determines the initial receptor. AUTOFIT does not
prompt the user for size fraction because it will do both size fractions.
The AUTOFIT sequence next invokes the CMB command and queries the user for
the analysis method (ridge regression, effective variance, and intercept
options). The method chosen will be used for all receptors and both size
fractions in the AUTOFIT series.
The AUTOFIT series begins with the fine fraction of the chosen initial
receptor site and date. Then the coarse fraction is analyzed. These
analyses are repeated on the next receptor site and date of the Receptor
Concentration Data File. The series proceeds in this manner to the end of
the file unless terminated by the user.
The user selects the initial receptor site and date according to
prompts and replies identical to those under the SELECT command. Then
the user chooses the analytical method according to prompts and replies
identical to those under the CMB command. There are no further prompts
for the rest of the AUTOFIT series except the following (see PDATA
command):
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
This prompt follows the displayed CMB coefficients. If the user
enters a null line by pressing RETURN, the program displays the calculated
concentrations in addition.
Unlike SELECT, entering a "C" under the AUTOFIT command will not
bring up the prompt "ENTER COMMAND." Instead, the program proceeds to the
47
-------
next analysis, which it also does after displaying the calculated concen-
trations. The AUTOFIT sequence can be terminated at this point by keying
in "STOP." The program then prompts "ENTER COMMAND."
If the user determines that he has mistakenly terminated the AUTOFIT
series, it is not necessary to start over. The command RESUME may be
entered immediately after the stop response to continue the AUTOFIT
series.
3.3.16 RESUME
The RESUME command should only be used in conjunction with the AUTOFIT
command. If the AUTOFIT sequence has been mistakenly terminated by a STOP
response, the RESUME command will continue the autofit sequence.
3.3.17 BACKOUT
The BACKOUT command allows the user to remove the effect of background
concentration from receptor concentration data. Subsequent CMB analyses
reflect the background adjustment. The BACKIN command reverses the BACKOUT
effects.
The BACKOUT command acts on the site/date receptor of the latest
SELECT (or AUTOFIT) command. The queries and replies for the BACKOUT
command are similar to those of the SELECT command:
ENTER BACKGROUND CMB SITE CODE: XXXXXXXXXXXX
The user should enter the CMB site code for the site chosen to repre-
sent background concentration. Subsequent queries identical with those of
the SELECT command ask for year and date. The user is not queried for
size fraction since both sizes are adjusted.
48
-------
3.3.18 BACKIN
The BACKIN command removes the background adjustment to receptor
concentrations applied by BACKOUT. The user should be aware that some
roundoff error may occur so that the BACKIN reversal of BACKOUT may not be
perfect.
3.3.19 PCOMP
The PCOMP command computes and prints the average contributions of
each source from a series of CMB analyses. The PCOMP command should follow
an AUTOFIT command or a series of SELECT-WRITE commands emulating on AUTOFIT
series.
A typical set of receptor concentration data could include a series of
sampling dates at one receptor station or a group of receptors in a target
area. The user would naturally want to compute the "average" CMB results
in these cases. This is the purpose of PCOMP.
For PCOMP to work properly it must be preceded by a series of SELECT
and WRITE commands for both fine and coarse fractions of each desired
receptor concentration data set. The AUTOFIT does this automatically.
Careful use of the SELECT and WRITE commands achieve the same end. However,
whenever possible, the AUTOFIT command should be used to avoid errors.
The PCOMP command computes the average contribution from all complete
CMB analyses following the previous PCOMP command or the initialization
phase, whichever is later. The complete CMB analyses are those that
included both fine and coarse analyses each followed by the WRITE command.
The output includes the labeled CMB coefficients from the averaged
analyses, the averages according to sources, and the standard deviations.
For each source, the mean and standard deviation are computed directly
49
-------
from the solutions for the different cases. A "solution" here is the
calculated contribution to the ambient aerosol of a particular source.
The means and standard deviations are not adjusted to account for the
correlated error effect discussed in Section 2.
The PCOMP command gives one prompt to the user:
OUTPUT WILL GO TO HARDCOPY
DO YOU WANT IT DISPLAYED AT YOUR TERMINAL INSTEAD?
The user response "NO" will result in the output being written on the
hardcopy. A "YES" brings the display to the user's terminal.
3.3.20 EXIT
The EXIT command terminates the interactive session. The user is
returned to the time sharing system.
3.4 Detailed Formats of Data Files
The various input data files must be coded according to the specific
formats presented here. The formats are presented as card images. How-
ever, the data should be on files accessible in interactive mode. For the
receptor concentrations the data must be on a unit that responds to a
FORTRAN REWIND. The other input data files are read only once.
Each data file line format is presented. Except for the receptor
concentration data file, each data file includes only one card type.
Since the CMS program is written in FORTRAN, the FORTRAN format of each
item is given for the convenience of the user in determining acceptable
entries.
In the following, the FORTRAN formats A, I, and F are referred to.
The A format is used to read alphanumeric information. The integer
50
-------
following the A, e.g., the 8 in A8, gives the width of the field. The I
format is used to read integers. For example, if the format is 14, the in-
teger must be right-justified in a 4-character-wide field. Decimal points
are not used with I format. The F format is used to read "floating-point,"
or arbitrary real numbers. The format F8.6 indicates that the field is 8
characters wide, and the rightmost 6 characters occur after the decimal
point. For example, the number 12345678 would be read as 12.345678. More
commonly, however, the decimal point is supplied. In this case the number
can appear anywhere in the field.
All of the following numbers would be read as 12.3 if format F8.6 was
used:
12.3
12.3
where the underscores indicate blanks.
Source Names and Codes
Column
1-2
3-4
5-12
13-80
Format
12
A8
Contents
Source code.
Skipped.
Source Name .
Skipped. May contain user comment.
Column
1-2
3-4
5-12
13-80
Format
12
A8
Species Names and Codes
Contents
Species Code.
Skipped.
Species Name.
Skipped. May contain user comment.
51
-------
The fine and coarse fraction source signature matrices use the same
line format.
Column
1-2
3-4
5-6
7-8
9-16
17-18
19-26
27-80
Format
12
-
12
-
F8.6
-
F8.6
-
Source Signature Matrix
Contents
Source Code.
Skipped.
Species Code.
Skipped.
Fraction of Source for Species.
Skipped
Standard Error of Fraction.
Skipped. May contain user comment.
A species code of '01' should not be used. Species code one is re-
served for total aerosol. The program automatically generates the source
signature vector for total aerosol.
The Receptor Concentration Data Set includes two card types. The card
type '03' identifies the receptor. There is only one card type '03' per
sampling date for each receptor location. The card type '30' gives the
actual concentration data. There is a card type '30' for each species. The
card type '30' lines follow directly after the corresponding card type '03-'
The concentration data may consist of fine and coarse fractions or of
the fine fraction and total. This data type is indicated on the card type
'03.'
Species code '01' is reserved for total aerosol. The total aerosol
data are used in selecting the "best" ridge regression solution.
52
-------
Receptor Concentration Data — Card Type 03 (Identification)
Column
1-2
3
4-15
16
17-18
19-22
23
24-25
26
27-28
29-32
33-34
35-80
Format
12
-
3A4
-
A2
A4
-
12
-
12
-
12
-
Contents
'03' — card type.
Skipped.
Receptor identification.
Skipped .
Year.
Month and Day (MMDD) .
Skipped.
Duration of Sample (hours) .
Skipped.
Start hour of Sample.
Skipped.
Data Type.
'12' — Fine and Coarse
'13' — Fine and Total
Must be blank.
Column
1-2
3
4-15
16
17-18
19-22
23
24-25
26
27-28
29-32
33-34
35-36
37-45
46-47
48-56
57-58
59-67
68-69
70-78
79-80
Receptor
Format
12
-
3A4
-
A2
A4
-
12
-
12
-
12
-
F9.4
-
F9.4
-
F9.4
-
F9.4
—
Concentration Data — Card Type 30
Contents
'30'— card type.
Skipped .
Receptor identification.
Skipped.
Year.
Month and Day (MMDD) .
Skipped.
Duration of Sample (hours) .
Skipped.
Start hour of Sample.
Skipped .
Species Code.
Skipped.
Fine Fraction Concentration.
Skipped.
Standard Error of Fine Fraction.
Skipped.
Coarse Fraction/Total Concentration.
Skipped.
Standard Error of Coarse Fraction/Total.
Skipped.
53
-------
The user followup data file is produced by the program and i-.- not
coded by the user. The format of the user followup data file is documented
here for those who wish to make use of it.
The user followup data file includes three card types. Card types
'03' and '30' are the same as Receptor Concentration Data Set card types
'03' and '30'. Card type '40' gives the species concentration contributed
by each source.
Recall that species 01 is the aerosol mass concentration for the size
fraction being considered. The card type '40' entries for species 01 give
the total aerosol contributions for their size fraction for the different
sources.
Card types '03' and '30' are written whenever a receptor site and
date are selected by the SELECT or AUTOFIT commands. Card type '40' is
written by the WRITE command when both fine and coarse size fractions
have been analyzed in succession.
Column
1-2
3
4-15
16
17-18
19-22
23
24-25
26
27-28
29
30-31
32
User
Format
12
-
3A4
-
A2
A4
-
12
-
12
-
12
-
Followup Data — Card Type 40
Contents
'40' — card type.
Skipped .
Receptor identification.
Skipped .
Year.
Month and Day (MMDD) .
Skipped.
Duration of Sample (hours) .
Skipped .
Start hour of Sample.
Skipped .
Source Code.
Skipped.
54
-------
User Followup Data — Card Type 40 (continued)
Column
33-34
35-36
37-45
46-47
48-56
57-58
59-67
68-69
70-78
79-80
Format
12
-
F9.4
-
F9.4
-
F9.4
-
F9.4
-
Contents
Species Code.
Skipped.
Fine Fraction
Skipped.
Uncertainty of
Skipped.
Contribution (yg/m3)
Fine Fraction.
Coarse Fraction Contribution.
Skipped .
Uncertainty of
Skipped.
Coarse Fraction.
3.5 Command Reference Summary
This section gives a short description of the purpose and results of
each command. This summary is intended to serve as a quick and ready-
reference for the user. Commands must be spelled out in full. However, a
response to a query may be abbreviated by the first letter unless a
numerical value or alphanumeric code is requested. Command details are
given in Section 3.2.
HELP—This command lists all the commands with a brief statement of
the purpose of each one.
SELECT—This command is used to select the receptor data set for the
current CMB analysis. SELECT also automatically invokes CMB and PDATA. The
user is prompted for the site, year, date and size fraction.
CMB—This command is used to determine the specific method options and
to perform the CMB analysis. CMB also automatically invokes PDATA. Method
options with CMB are ridge regression/least squares, effective variance
method/no effective variances, and fitted model intercept/zero intercept.
These options may be exercised in any combination.
55
-------
PDATA—This command displays the current analysis solution at the
user's terminal. The output consists of the solution coefficients and
standard errors. The least squares coefficients, standard errors, and
variance inflation factors are also displayed. Optionally, the receptor
data and predicted values may be displayed for all species.
WRITE—This command writes results to hardcopy files. The WRITE
command should follow the SELECT or CMB commands.
AE—This command is used to add species to the current fitting set.
The program prompts the user for species to add.
DE—This command is used to delete species from the current fitting
set. The program prompts the user for species to delete.
AS—This command is used to add sources to the current fitting set.
The program prompts the user for sources to add.
DS—This command is used to delete sources from the current fitting
set. The program prompts the user for sources to delete.
AUTOFIT—This command is used to perform automatically CMB analyses on
a series of sites or dates (for both fine and coarse fractions) . The
AUTOFIT command saves time and prevents errors by eliminating repetitive
entries. The AUTOFIT command is equivalent to a series of SELECT (CMB,
PDATA) and WRITE commands.
RESUME—This command is used in conjunction with AUTOFIT to resume the
autofit sequence.
BACKOUT—This command is used to remove the effect of background
concentrations from receptor data. This command should be invoked after
SELECT. The program prompts the user for site, year and date.
56
-------
BACKIN—This command is used to cancel the effect of the BACKOUT
command.
PMATRIX—This command is used to display the source signature matrix,
the receptor data, species code list, and source code list. The program
prompts the user for the items to display. Optionally, the output may go
to the user's terminal or to hardcopy.
PSOLN—This command is used to display all ridge solutions, not merely
the "best" one. Optionally, the output may go to the user's terminal or to
hardcopy.
PINFO—This command displays the current problem status giving site,
date, size, and background information. It is primarily useful to users of
CRT terminals.
PCOMP—This command is used to compute and print the averages and
standard deviations of a series of CMB analyses. Optionally, the output
may go to the user's terminal or to hardcopy.
EXIT—This command terminates the CMB program processing.
3.6. Computer System Considerations
Certain tasks relevant to using the CMB program are contingent on the
conventions of the particular computer being used. These include compila-
tion and subprogram linkage, file access, and program execution. This
section gives information pertinent to computer dependency. The emphasis
of this information is oriented towards the EXEC 8 system of the Univac
1100 series, which was used in the program development. The requirements
of IBM TSO are given a shorter discussion. The command names and syntax
vary, but the tasks that must be performed are similar regardless of
computer systems.
57
-------
The physical characteristics of the eleven files accessed by the CMB
program are as follows:
FORTRAN
Unit # Input/Output Record or Line Length
2 Input 80
3 Input 80
5 Input 72 (Terminal)
6 Output 72 to 132 (Terminal)
7 Input 80
8 Input 80
9 Output 121 (Print File)
10 Input/Output 80 (Scratch Data File)
11 Output 133 (Print File)
12 Output 80
13 Input 80
The CMB program consists of a main program and five subprograms. All
programs are written in FORTRAN. These programs are the following:
Program Purpose
MAIN Initialization and command sequence
FETCH Read receptor concentration data
CMBR Perform CMB analysis
PERC Compute percentages and variances
COARS Compute coarse fraction and variances
INV1 Matrix inversion
The program MAIN calls FETCH, CMBR, and PERC. FETCH calls COARS;
CMBR calls INV1.
3.6.1 Uuivac System Conventions
This section describes one approach to using the CMB program on a
Univac 1100 series computer. Other approaches are feasible. Tf the direc-
tions given here are followed, the CMB program can be invoked by a statement
like the following:
@ADD CMB-RADPRG.HCS-FILE
58
-------
Similarly, the hardcopy files may be printed by a statement like the
following:
x
@ADD CMB-RADPRG.HCS-PRT
The following steps prepare the source code for execution. These
steps need be performed only once if the result is saved.
@FTN,U CMB-RADPRG.MAIN,CMB-RADPRG.MAIN
@FTN,U CMB-RADPRG.FETCH,CMB-RADPRG.FETCH
@FTN,U CMB-RADPRG.CMBR,CMB-RADPRG.CMBR
@FTN,U CMB-RADPRG.PERG,CMB-RADPRG.PERC
@FTN,U CMB-RADPRG.COARS, CMB-RADPRG.COARS
@FTN,U CMB-RADPRG.INV1,CMB-RADPRG.INV1
@MAP,I CMB-RADPRG.MAP,CMB-RADPRG.CMB
IN CMB-RADPRG.MAIN,.CMBR,.COARS,.FETCH,.PERC,.INV1
END
The input data files must contain the data for the analyses. If the
data are punched on cards, the cards may be loaded to the files using the
@ELT command, as in the following example:
@ELT CMB-RADPRG.HCS-SORF
—DATA CARDS FOR FINE SOURCE SIGNATURE—
@ELT CMB-RADPRG.HCS-SORC
—DATA CARDS FOR COARSE SOURCE SIGNATURE—
@ELT CMB-RADPRG.HCS-DATA
—DATA CARDS FOR RECEPTOR CONCENTRATION DATA—
@ELT CMB-RADPRG.HCS-PONM
—DATA CARDS FOR SPECIES NAMES & CODES —
@ELT CMB-RADPRG.HCS-SONM
—DATA CARDS FOR SOURCE NAMES 6. CODES—
59
-------
Alternatively, the data may be entered directly onto the file in an
interactive session using the editor:
@ED,I CMB-RADPRG.HCS-SORF
—DATA LINES FOR FINE SOURCE SIGNATURE--
ETC.
The following command sequence will define the files and begin
execution of the CMB program. For convenience, these commands are best
placed on a file element (using the editor) and brought into the runstream
with an @ADD command. The approach given here assigns temporary SDF files,
copies the permanent data onto the temporary files, sets up the FORTRAN
unit numbers, and executes the CMB program.
@ASG,A CMB-RADPRG.
@FREE CMB-RADPRG.
@ASG,T TEMP.
@ASG,T MAGSTO.
@ASG,T SUMMRY.
@ASG,T CMBOUT.
@ASG,T HCS-PONM.
@ASG,T HCS-SORC.
@ASG,T HCS-SORF.
@ASG,T HCS-DATA.
@ASG,T HCS-SONM.
@COPY,I CMB-RADPRG.HCS-PONM,HCS-PONM.
@COPY,I CMB-RADPRG.HCS-SORC.HCS-SORC.
@COPY,I CMB-RADPRG.HCS-SORF,HCS-SORF.
@COPY,I CMB-RADPRG.HCS-DATA,HCS-DATA.
@COPY,I CMB-RADPRG.HCS-SONM,HCS-SONM.
@USE 2.,HCS-SONM.
@USE 3.,HCS-PONM.
(?USE 7., HCS-SORF.
60
-------
-------
These lines would be the content of the file CMB-RADPRG.HCS-FILE used
in the @ADD statement given at the beginning of this subsection.
Alternatively, the results may be saved permanently for later printing
on the system printer:
@COPY,I CMBOUT.,CMB-RADPRG.CMBOUT
(?COPY, I SUMMRY . , CMB-RADPRG. SUMMRY
@COPY,I MAGSTO.,CMB-RADPRG.MAGSTO
3.6.2 IBM ISO System Considerations
The CMB program was developed on a Univac 1100 series computer. Never-
theless, every conscious effort was made to render the code compatible with
IBM FORTRAN.
This section will briefly discuss the IBM ISO conventions needed to
execute the CMB program.
The CMB programs can be compiled using the TSO FORT command or one of
the FORTRAN batch compilers. The programs can be linked into a load module
using the TSO LINK command or the batch linkage editor.
The input data files may be loaded from cards using the IBM utility
IEBGENER. Alternatively, the lines may be entered directly using the TSO
EDIT command.
The data files may be assigned FORTRAN unit numbers using the TSO
ALLOCATE command, as in the following example:
ALLOC DA(CMB.HCS.PONM.DATA) F(FT03F001)
62
-------
The program can be executed using the ISO CALL command, as in the
following example:
CALL CMB.PD S.LOAD(CMB)
The ALLOCATE commands and the CALL command can be organized into a
CLIST so that a single EXEC command will begin the CMB program execution.
The hardcopy prints may be displayed at the user's terminal using
the TSO EDIT LIST command or sent to the system printer using the IBM
utility PRTPCH.
63
-------
SECTION 4
EXAMPLE RUN
4. 1 Introduction
This section presents an example interactive session using the CMB
program. This run of the CMB program exercises all of the commands. The
purpose of this run is to illustrate the use of each command, and, as a
result, it does not necessarily follow the course that an actual analysis
might take. Moreover, the course of an actual analysis can vary signifi-
cantly, depending on the nature and size of the ambient receptor data set
and the objectives to be achieved by the analyses.
The example runs were executed on EPA's Univac 1100 series computer.
The program was invoked with the command
@ADD CMB-RADPRG.HCS-FTLE
as described in Section 3.6. The user's entries are preceded in the
session by the symbol ">".
4.2 Illustration of Commands
UOULD YOU LIKE TO LOOK AT THE SOURCE AND SPECIES LISTS7
TYPE YES OR NO.
The user replies "YES" so that he will geC a list of the source and species
numbers to refer to in the rest of the interactive session.
64
-------
>YE8
SOURCE SIGNATURE MATRICES ARE CODED AS FOLLOWS
SOURCES:
SOURCE tt SOURCE NAME
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
SOIL
RD DUST
SEA SALT
SLSH BRN
AUTO CAT
AUTO
JET AIR
RES OIL
DlSf OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRHNFR
CAL GYP
CEMENT
PET FCC
LSKILN
SEC S04
SPECIES:
SPECIES «
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
SPECIES NAME
TOTAL
AL
AS
BA
8R
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
S04
SB
SI
SN
SR
TI
65
-------
25
2b
T7
4. /
23
29
30
31
V
ZN
C
NA
N03
RB
SE
Next, the" user selects the sources using the source numbers in the
above list. The user keys a "RETURN" or null line for "INITIAL SOURCE 15'
to indicate that he has entered all of the desired sources. The program
displays a list of the selected sources with their source numbers for the
user's verification.
PLEASE INPUT INITIAL FITTING INFORMATION
INITIAL SOURCE 1: XX
>1
INITIAL SOURCE 2: XX
>2
INITIAL SOURCE 3: XX
>3
INITIAL SOURCE 4: XX
>5
INITIAL SOURCE 5: XX
>6
INITIAL SOURCE 6: XX
:.-?
INITIAL SOURCE 7: XX
,-9
INITIAL SOURCE 8: XX
>10
INITIAL SOURCE 9: XX
rl1
INITIAL SOURCE 10: XX
-12
INIflAL SOURCE 11: XX
." ! ,3
INITIAL SOURCE 12: XX
66
-------
INIHAL SOURCE 13: XX
:>1A
INITIAL SOURCE 14: XX
>17
INITIAL SOURCE 15: XX
t/
INITIAL SOURCES
1
2
3
5
6
7
9
10
11
12
13
14
U
17
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
JET AIR
DIST OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRMNFR
CEMENT
PET FCC
Next the user selects the species. This can be done individually as
above with the sources. However, the user wants to enter "all" of the
species. He keys a "-1" for "INITIAL SPECIES 1" and the first 21 (maximum
allowable) species are selected. The program displays a list of them for
the user's verification.
PLEASE INPUT SPECIES
INITIAL SPECIES 1: XX
>-1
INITIAL SPECIES
1
2
3
4
5
6
7
8
9
TOTAL
AL
AS
6A
BR
CA
CD
CL
CO
-------
10
11
12
13
14
15
16
17
18
19
20
21
CR
CU
FE
HG
K
m
NI
P
PB
S04
SB
SI
Next the program asks the user if the selected sources and species are
correct. The user is satisfied with his choices (for the time being) and
replies "YES." Then the program notes that the current size fraction is
fine (rather than coarse). The session now enters the command phase.
ARE INITIAL SOURCES AND SPECIES CORRECT?
>YES
THE "HELP" COMMAND LISTS COMMANDS
SIZE IS FINE
ENTER COMMAND
The user decides to give the HELP command to get a list of the
commands and their uses.
••HELP
HELP-LISTS THESE COMMANDS
DATA ACCESS AND SEQUENCING
AUTOFIT-3EQUENCES AUTOMATICALLY TO NEXT DATA SET
RESUME-RESUME AUTOFIT
SELECT-SELECT DATA SET FOR CMB
SIZE-CHANGE SIZE FRACTION
EXIT-CLOSE FILES AND LEAVE
68
-------
CHB OPERATIONS
AE-ADD A SPECIES TO THE FIT
DE-DELETE A SPECIES FROM THE FIT
AS-ADD A SOURCE 10 THE FIT
DS-DELETE A SOURCE FROM THE FIT
CMB-PERFQRH CMS: OPTIONS ARE RIDGE * EFFECTIVE VARIANCE
SCREEN DISPLAY
PINFO-PRINT CURRENT STATUS ON SCREEN
PDATA-PRINT CURRENT CUB RESULTS ON SCREEN
PMATRIX-PRINT SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS,
OR SOURCE AND SPECIES CODE LISTS
PSDLH-PRINT ALL RIDGE SOLUTIONS
PCOhF'-PRINT COMPUTED AVERAGES OF CMB SERIES
DATA STORAGE
URITE-URITE PRESENT CHB RESULTS TO FULL PRINTOUT,
SUMMARY.AND USER FOLLOUUP STORAGE FILES
BACKGROUND SITE OPERATIONS
BACKOUT-SELECT AND SUBTRACT A BACKGROUND DAI A SET
BACKIN-ELIHINATE CURRENT BACKGROUND SUBTRACTION
SELECT SHOULD NORHALLY BE THE FIRST COMMAND
ENTER COMMAND
The SELECT command will bring in receptor concentration data.
>SELECF
ENTER DESIRED C«B SITE CODE: XXXXXXXXXXXX
If the user does not know how the receptor sites are coded, he can
reply with a question mark or any other incorrect response. Then the
program will list the sites for him:
ENTER YEAR: YY
s
ENTER DATE: MMDD
INPUT DESIRED SIZE FRACTION: (FINE OR COARSE)
69
-------
UATA SEARCH BEGUN FOR
SITE: ? YEAR: DftTE:
SITE: URBAN CORE rEAR: 81 DATE: 0229 FC
SITE: BACKGROUND rEAR: 81 DATE: 0229 FC
DATA SET NOT FOUND FOR
SITE: ? YEAR: BATE:
USE ONE OF THOSE LISTED ABOVE
ENTER COMMAND
Now the user knows what receptor site data are available. He gives the
SELECT command again, choosing the Urban Core site. The sites and dates
available depend on the data in the Receptor Concentration Data File de-
scribed in Section 3. The data available for this interactive session are
given in Appendix C.
^SELECT
ENTER DESIRED CHB SITE CODE: XXXXXXXXXXXX
XIRBAN CORE
ENTER YEAR: YY
>81
ENTER DATE: MMUD
X>229
[NPUT DESIRED SIZE FRACTION:(FINE OR COARSE)
•'-FINE
DATA SEARCH BEGUN FOR
SITE: URBAN CORE YEAR: 81 DATE: 0229
SITE: URBAN CORE YEAR: 81 DATE: 0229 FC
The program now automatically moves into the CMB command.
70
-------
DO YOU UANT RIDGE SOLUTIONS?
CMB UILL SELECT THE BEST RIDGE SOLUTION
DID YOU UANT TO SEE A SUMMARY OF THE OTHERS?
>NQ
IN THE SELECTION OF THE BEST RIDGE SOLUTION,
UHAT WEIGHT DO YOU UANT? (DEFAULTS .0)
';•
UEIGHT IS 1.000000000
BO YOU UANT TO USE EFFECTIVE VARIANCES?
>NO
DO YOU UANT TO INCLUDE THE INTERCEPT IN THE MODEL'
The user selects the method options by replying appropriately to the
questions asked. The summary of the ridge solutions is illustrated later in
this session. The default weight is selected by keying a RETURN or null line.
Keying in "1.0" would have the same effect. Note that if a weight is given,
the decimal point must be included.
The program now automatically enters the command PDATA.
CMB SITE: URB*
SAMPLE DURATIC
n T TI f* r
1
2
3
5
6
7
9
10
1 I
12
13
14
16
17
IMVIUE
R-SQUARE: .{
INTERCEPT
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
JET AIR
DIST OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRMNFR
CEMENT
PET FCC
TOTAL:
iN CORE
IN: 12 UITH
• REGRESSIO
)991 RIDG
i ir* / u 1
-
9
4
3
1
-
6
7
34
uu/ no
.004+-
.448+-
.011+-
.313+-
.566+-
.013+-
.520+-
.944+-
.295+-
.931+-
.095+-
.028+-
.085+-
.487+-
.900+-
.824+-
MEASURED CONCENTRATION F
35.00000+-
7.
000 / 7
YEAR: 81 DA
START HOUR: 7
Kl- _».
E K=
1.
,
,
9.
,
44.
5.
9
1.
2.
.
.
2.
6.
48.
.050
005
089
858
138
073
957
126
510
361
899
467
190
145
443
591
266
TE
Bi
: 0229 FRACTION
ACKGROUND: NO EFF
WEIGHTED LEAST SQUAR
R-SQUARE: .9139
1 If* /U '7 t t T C
uo/ no
.004+- .008
-6
-
-6
4
3
7
12
2
-
7
9
35
.760+-
.878+-
.572+-
.683+-
.790+-
.542+-
.215+-
.645+-
.149+-
.912+-
.033+-
.126+-
.071+-
.716+-
.000+-
53.
6.
.
80.
1.
58.
18.
86.
16.
24.
3.
1.
27.
72.
50.
326
047
869
143
135
707
067
973
839
370
765
834
846
098
671
1333.
86.
128.
113.
3.
1 .
34.
1090.
139.
219.
553.
136.
237.
198.
: FINE
VAR: NO
re -.
to
67?
913
630
999
290
941
277
330
242
996
312
655
782
829
INE/CQARSE/TOTAL:
5.00000+- /.OOO/
110.00000+-
9.399
71
-------
This table presents the fitted coefficients and their standard errors
for each source. Two sets of coefficients are given: the solution on the
left is the "best" ridge solution (k = 0.050) and the solution on the right
is the weighted least squares solution (k = 0.0).
The solutions consist of coefficient estimates with standard errors for
each source. Variance inflation factors are given at the far right. At the
bottom the calculated total aerosol is given for each solution. The
measured total aerosol is also given for each size fraction.
Due to the complexity of the set of sources used and possibly also to
uncertainties in the source signatures, both solutions contain large random
errors for some sources. The large variance inflation factors for some
sources indicate that strong multicollinearities exist. Notice, however,
that the ridge solution has greatly reduced standard errors for the sources
with large variance inflation factors. Also, the ridge solution has no
significant negatives, although the weighted least squares solution has
negatives with large magnitudes. This is typical in problems affected by
multicollinearity.
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
The user can now proceed to a new command by entering a "C." However,
the user wants to examine the calculated concentrations. He presses the
return key to get this table:
72
-------
1 TOTAL
2 AL
3 AS
4 BA
5 BR
6 CA
7 CD
8 CL
9 CO
10 CR
11 CU
12 FE
13 H6
14 K
15 HH
16 NI
17 P
18 PB
19 S04
20 SB
21 SI
22 SN
23 SR
24 TI
25 V
26 ZN
27 C
28 NA
29 N03
30 RB
31 SE
ntMo
* 35.000+- 7
* .088+-
* .036+-
* .149+-
* .344+-
* .441+-
* .004+-
* ,051+-
* .003+-
* <
••K .013+-
* .208+-
* <
•* .145+-
* .022+-
* .006+-
* .119+-
* 1.422+-
* 13.130+-
* .003+-
* .277+-
<
.003+-
.X
<
.042+-
M <
M <.
M <
M <
M <
.000
.062
.007
.014
.017
.023
.002
.007
.002
.004
.002
.011
.003
.009
.004
.001
.017
.068
.681
.002
.031
.002
.001
.014
.008
.003
.001
.001
.001
.001
.001
cftLC RATIg
34.824+-4B.266 .99+- 1.39 TOT
.033+- .004 .38+- .27 AL
ENTER COMMAND
.004+-
.005+-
.334+-
.441+-
.006+-
.060+-
.004+-
.005+-
.009+-
.203+-
.004+-
.148+-
.000
.000
.082
.082
.000
.069
.000
.000
.001
.029
.000
.033
1
I
1
1
2
.12+-
.03+-
.97+-
.00+-
.55+-
.17+-
.50+-
.58+-
.71+-
.97+-
.00+-
.90+-
.02
.00
.24
.19
.78
1.37
1.00
5.16
.12
.15
.00
.20
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
.004+- .000
.870+- .212
13.027+- 1.832
.023+- .004 1.06+- .26 MM
.006+- .000 1.04+- .18 NI
.04+- .01 P
.61+- .15 PB
.99+- .15 304
.005+- .000 1.68+- 1.12 SB
.305+- .046 1.10+- .21 SI
004 20.40+-40.99 SN
000 1.54+- .51 SR
1.09+- 2.18
.001
.000 2.29+- 9.15
.21
.020+-
.005+-
.008+-
.005+-
.046+- .008 1.08+
10.518+- 1.383
-.094+- .024
.019+- .003
.004+- .000
.004+- .000
TI
V
ZN
.00+- .00 C
.00+- .00 NA
.00+- .00 N03
.00+- .00 RB
.00+- .00 SE
For each species available the table above gives the measured concen-
tration, the calculated concentration, and the ratio of the calculated to
the measured concentration. Each quantity is accompanied by its standard
error.
Those species that were included in the fitting process are marked with
an asterisk (*). Those species that are marked with an "M" do not have
receptor concentration data.
73
-------
The user has the option to change the selection of sources and species
and reperform the analysis. The effect of deleting undetected species Cr
and Hg (10 and 13) and adding the detected but unused species Sr and Zn
(23 and 26) could be investigated. In general, however, species which
serve as key tracers for important sources should be included even if those
species have small ambient concentrations. A tracer could be a species
which constituted a small percentage of the aerosols from a given source
but which was virtually unique to that source. Sources with negative
coefficients are candidates for removal. Also, sources with standard errors
many times larger than their coefficients are potential deletions. The in-
dication is that the contribution of such sources is zero within random
variability. The large SOi* measured concentration indicates that sulfate
sources are important. The secondary sulfate source (19) is a potential
addition.
Upon examining his first trial the user determines that he wishes to
change some of the fitting species and sources. He uses the DE command
to delete species, the AE command to add species, the DS command to delete
sources, and the AS command to add sources. In each case, the user termi-
nates the command with a RETURN.
SIZE IS FINE
INPUT CODE OF DELETED SPECIES
no
INPUT CODE OF DELETED SPECIES
XI 3
INPUT CODE OF DELETED SPECIES
ENTER COflhAND
,-AE
SU'E IS FINE
INPIJI CODE OF ADDED SPECIES
CODE l)F ADDED SPECIES
74
-------
INPUT CODE UF ADDED SPECIES
ENTER COMMAND
>DS
SIZE IS FINE
INPUT CODE OF DELETED SOURCE
>12
INPUT CODE OF DELETED SOURCE
'>7
INPUT CODE OF DELETED SOURCE
,/•
ENTER COKMAND
>AS
SIZE IS FINE
INPUT CODE OF ADDED SOURCE
>19
INPUT CODE OF ADDED SOURCE
.,>
ENTER COMMAND
Now the user is ready to rerun the CMB analysis. It is not necessary
to use SELECT again since the data have been already brought in for the
desired site. Therefore, the user gives the CMB command.
,-CMB
DO YOU WANT RIDGE SOLUTIONS?
>YES
CMB yiLL SELECT THE BEST RIDGE SOLUTION
DO YOU UANT TO SEE A SUMMARY OF THE OTHERS?
>NO
IN THE SELECTION OF THE BEST RIDGE SOLUTION,
WHAT WEIGHT DO YOU UANT? (DEFAULTS.0)
WEIGHT IS 1.000000000
DO YOU WANT TO USE EFFECTIVE VARIANCES?
;-NG
DO YOU WANT TO INCLUDE THE INTERCEPT IN THE MODEL?
75
-------
CrtB SITE: URBAN
SAMPLE DURATION
r. | r, r* L
ENTER COMMAND
This time the user gives a "C" and no receptor concentrations are
displayed.
The user is now ready to look at the coarse fraction. The SIZE command
switches the current size to coarse. Then the user adjusts his selection of
sources. The species and sources for the coarse size fraction were not
affected by the adjustments made for the fine fraction earlier.
76
-------
bl/.E
SIZE IS COARSE
ENTER COMMAND
>DE
SIZE IS COARSE
INPUT CODE OF DELETED SPECIES
X5
INPUT CODE OF DELETED SPECIES
>?
INPUT CODE OF DELETED SPECIES
,-13
INPUT CODE OF DELETED SPECIES
>16
INPUT CODE OF DELETED SPECIES
>19
INPUT CODE OF DELETED SPECIES
>20
INPUT CODE OF DELETED SPECIES
ENTER COMMAND
>AE
SIZE IS COARSE
INPUT CODE OF ADDED SPECIES
"<23
INPUT CODE OF ADDED SPECIES
>24
INPUT CODE OF ADDED SPECIES
>26
INPUT CODE OF ADDED SPECIES
ENTER COMMAND
:-DS
SIZE IS COARSE
INPUT CODE OF DELETED SOURCE
;-','
INPUT CODE OF DELETED SOURCE
:•1,7
INPUT CODE OF DELETED SOURCE
ENTER COMMAND
77
-------
Now the user asks for a CMB analysis of the coarse fraction.
DO YOU UANT RIDGE SOLUTIONS?
.'SAME
CrtB SITE: URBAN CORE YEAR: 81
SAflPLE DURATION: 12 UI TH START HOUR:
RIDGE REGRESSION
R-SQUARE: .6151
SOURCE
INTERCEPT .011+-
I SOIL 36.445+-
2 RD DUST 13.578+-
3 SEA SALT .177+-
5 AUTO CAT 10.339+-
6 AUTO .722+-
11 COAL 4.532+-
12 KRAFT RB 2.229+-
16 CEMENT 13.874+-
TOTAL: 81.908+-
m
E K= .300
.005
18.752
6.185
.491
13.655
.275
3.137
3.455
11.324
25.267
wt ioni 1
R-Si
IIP / U'
U(j/n.
.023+-
77.894+-
18.281+-
.365+-
-5! .096+-
1 .180 + -
11 .906+-
-.031 +-
23.364+-
81.834+-
DATE: 0229 FRACTION: COARSE
7 BACKGROUND: NO EFF VAR: NO
UEI6HTED LEAST SQUARES
ARE: .7702
.010
81.347
13.310
.771
63.596
.395
7.909
6.302
20.317
43.046
,178
,134
,223
.754
.054
.829
.564
.550
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
35.00000+- 7.000/ 75.00000+- 7.000/ 110.00000+- 9.899
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMrtAND
••€
ENTER COMMAND
Note that the user replied "SAME" to the question "DO YOU WANT RIDGE
SOLUTIONS?" The user wishes to use the same analysis method as in the
previous CMB command. Thus, the user has asked for ridge solutions, no
ridge summary, a selection weight of 1.0, no effective variances, and an
intercept included in the model. Using the SAME reply can eliminate repeti-
tive replies if the user chooses to use a consistent analysis method.
The user issues a WRITE command to record the output on hardcopy. The
hardcopy output from this session is given in Appendix C. The hardcopy out-
put includes tables similar to the source coefficient table and the
78
-------
calculated concentration table. Note that the latter will be included on
the hardcopy even though it was not printed at the user's terminal.
>URITE
URITTEN
ENTER COMMAND
The user now would like to see the same analysis of the coarse fraction
but with "background" removed. He gives the BACKOUT command and selects .
a background concentration receptor.
>BACKOUT
ENTER BACKGROUND CHB SITE CODE: XXXXXXXXXXXX
>BACKGROUND
ENTER YEAR: YY
ENTER DATE: HMDD
>0229
SITE: URBAN CORE YEAR: 81 HATE: 0229 FC
SITE: BACKGROUND YEAR: 81 DATE: 0229 FC
BACKOUT COHPLETE
ENTER COMMAND
Now the user gives the CMS command to perform the analysis with back-
ground removed.
.5CMB
DO YOU UANT RIDGE SOLUflONS?
.'•SAME
79
-------
CMB SITE: URBAN CORE YEAR: 81
SAMPLE DURATION: 12 UITH START HOUR:
RIDGE REGRESSION
R-SQUAKE: .4954 RIDGE K= .800
INTERCEPT
1 SOIL
2 Rll DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
11 COAL
12 KRAFT RB
16 CEMENT
TOTAL:
uu/ n,
.010+-
21.685+-
8.110+-
.192+-
6.265+-
.256+-
3.435+-
2.065+-
8.208+-
50.225+-
.004
13.296
4.089
.300
9.969
.207
2.114
2.730
7.614
21.945
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
7.00000+- 7.6167 38.00000+- 8.062/ 45.00000+- 11.091
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
ENTER COMMAND
DATE: 022? FRACTION: COARSE
7 BACKGROUND: YES EFF VAR: NO
WEIGHTED LEAST SQUARES
R-SQUARE; .6667
UG/M3 VIF
.021+- .008
56.462+- 63.333
9.882
.543
o 1 • / .<.*.
.313
6.250
5.670
14.487
37.577
9.703+-
.428+-
-48.147+-
.350+-
9.835+-
-.074+-
14.206+-
42.783+-
6.733
2.465
1.423
7.508
1.059
3.414
1 .772
1 .575
Note that the user again uses the reply SAME to eliminate repetition
in asking for the same analysis method as previously.
The BACKIN command removes the effect of the BACKOUT command. This
may be desired if a reanalysis of the raw site data is desired or if the
user would like to try another background site.
>BACKIN
BACKIN COMPLETE
ENTER COMMAND
The next command given by the user is AUTOFIT, so the BACKIN command
was not actually necessary in this case. The AUTOFIT command will sequence
through both size fractions of a selected receptor and all subsequent re-
ceptors on the Receptor Concentration Data File. It also writes the results
to the hardcopy.
80
-------
ENTER DESIRED CKB SITE CODE: XXXXXXXXXXXX
>URBAN CORE
ENTER YEAR: YY
ENTER DATE: M«DD
>0229
DATA SEARCH BEGUN FOR
SITE: URBAN CORE
YEAR: 81 DATE: 0229
SITE: URBAN CORE YEAR: 81 DATE: 0229
BO YOU WANT RIB6E SOLUTIONS?
>YES
CMB UILL SELECT THE BEST RID6E SOLUTION
DO YOU UANT TO SEE A SUHMARY OF THE OTHERS?
>NO
IN THE SELECTION OF THE BEST RIDGE SOLUTION,
WHAT UEIGHT DO YOU UANT? (DEFAULTS,0)
>
UEIGHT IS 1.000000000
DO YOU UANT TO USE EFFECTIVE VARIANCES?
>NO-'
NO
DO YOU UANT TO INCLUDE THE INTERCEPT IN THE MODEL?
>YES
FC
CMB SITE: URBAN CORE YEAR: 81
SAMPLE DURATION! 12 UITH START HOUR:
RID6E REGRESSION
R-SQUARE: .9179 RIDGE K= .010
DATE: 0229 FRACTION: FINE
7 BACKGROUND: NO EFF VAR: NO
UEIGHTED LEAST SQUARES
R-SQUARE: .9200
auuKLt
INTERCEPT
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
9 DIST OIL
10 RES GAS
11 COAL
I.J ELARCFRN
14 FERRMNFR
16 CEMENT
'I? PET FCC
19 SEC S04
TOTAL:
uu/no
.004+-
-.509+-
-.024+-
-.446+-
7.377+-
4.635+-
3.596+-
1.015+-
1.771+-
.076+-
.061+-
7.403+-
8.221+-
-.187+-
32.991+-
.004
3.092
.923
.171
20.436
.941
6.614
5.311
1.348
.454
.250
2.919
12.252
11.012
15.162
uu/ riv
.005+-
-8.858+-
.891+-
-.533+-
-2.525+-
4.800+-
6.549+-
15.427+-
2.192+-
.048+-
.083+-
7.999+-
9.410+-
-3.353+-
32.136+-
j
.006
27.313
3.239
.265
58.257
.975
10.723
49.140
2.462
1.2.4V
.612
4.322
41.216
49.346
50.953
vii
432.869
29.638
14.234
79.398
2.881
14.366
413.108
9.374
71 .640
24.698
6.798
77.121
173.618
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
35.00000+- 7.000/ 75.00000+- 7.000/ 110.00000+- 9.899
81
-------
The user has entered information selecting the first site. He has also
chosen the analysis method. The method chosen will be used for the remain-
ing sites and size fractions for the AUTOFIT sequence.
PRESS RETURN TO CONTINUE OR ENTER C l:0k NEXT
The program pauses for a reply to the above prompt. However, a "C"
does not return the user to command mode. Instead the program follows the
above CMB analysis of the fine fraction of URBAN CORE with the CMB analysis
of the coarse fraction. Note also that hardcopy is "WRITTEN."
.•-C
WRITTEN
CMfi SITE: URBAN CORE YEAR: 81
SAMPLE DURATION: 12 WITH START HOUR:
RID6E REGRESSION
R-SQUARE: .6151 RIDGE K= .300
3UUKUC.
INTERCEPT
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
11 COAL
12 KRAFT RB
16 CEMENT
TOTAL:
uu/no
.011+-
36.445+-
13.578+-
.177+-
10.339+-
.722+-
4.532+-
2.229+-
1 3.874+-
81.908+-
.005
18.752
6.185
.491
13.655
.275
3.137
3.455
11.324
25.267
.0
77.8
18.2
.3
-51 .0
1 .1
11 .9
-.0
23.3
81. a
DATE: 0229 FRACTION: COARSE
7 BACKGROUND: NO EFF VAR: NO
UEIGHTEIi LEAST SQUARES
R-SQUARE: .7702
U6/H3 VIF
'23+- .010
94+- 81.347
:81+- 13.310
65+- .771
96+- 63.596
80+- .395
06+- 7.909
'31+- 6.302
64+- 20.317
:84+- 43.046
6.178
2.184
1 .223
6.754
1 .054
2.829
1 .564
1 .550
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
35.00000+- 7.000/ 75.00000+- 7.000/
!10.00000+- 9.899
82
-------
Note that the sources for the fine and coarse fractions are different.
The sources (and species) used are the selections currently in effect for
each size fraction. These were determined in the initialization phase and
by AS, AE, DS, and DE commands.
The user decides to display the calculated concentrations for the coarse
fraction of the URBAN CORE site.
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
— - — — . «.—_-— __ — - __..fj£fj^ -,_.._..__ ___ " LHL.U *• — -"—-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
SO 4
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
* 75.000+-
* 1.951+-
.004+-
* .371+-
* . 1 08+-
* 13.014+-
<
* .288+-
* .015+-
* .022+-
* .018+-
* 1.871+-
<
* .517+-
* .088+-
.003+-
* . 1 02+-
* .346+-
1.725+-
.004+-
* 8.652+-
.005+-
* .030+-
* .205+-
.:.
* .071+-
M
M <
M <
M <
M <
7.000
.422
.004
.023
.007
1.442
.001
.068
.003
.005
.002
.118
.002
.072
.008
.001
.040
.024
.939
.002
1.987
.002
.003
.037
.014
.006
.001
.001
.001
.001
.001
81.908+-25
1.125+-
.011+-
.011+-
.073+-
5.186+-
.011+-
.212+-
.011+-
.022+-
.018+-
1.204+-
.011+-
.347+-
.035+-
.018+-
.011+-
.171+-
6.525+- 1
.011+-
6.335+- 1
.011+-
.016+-
.181+-
.015+-
.054+-
8.1 29+- 1
.422+-
.014+-
.012+-
.011+-
.267
.229
.000
.000
.015
.842
.000
.031
.000
.003
.001
.205
.000
.070
.004
.001
.000
.038
.306
.000
.048
.000
.001
.035
.001
.008
.224
.063
.001
.000
.000
KH 1 1U
1.09+- .
.58+- .
2.76+- 2.
.03+- .
.67+- .
.40+- .
22.10+-44.
.74+- .
.74+- .
.99+- .
1.01+- .
.64+- .
11.05+-22.
.67+- .
.40+- .
5.83+- 1.
.11+- .
.49+- .
3.78+- 2.
2.76+- 1.
.73+- .
2. 21+- .
.54+- .
.38+- .
2. 10+- 4.
.76+- .
.00+- .
.00+- .
.00+- .
.00+- .
.00+- .
35
17
76
00
14
08
19
20
15
26
13
12
10
17
06
99
04
12
19
38
21
38
07
23
21
13
00
00
00
00
00
TOT
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
S04
SB
SI
SN
SR
ri
V
ZN
C
NA
N03
RB
SE
WRITTEN
UR
I TTEN
83
-------
Note that the program paused at the end of the table display. The
user gave a RETURN to continue processing. The program replied with two
"WRITTEN" messages, indicating that the coarse fraction results have been
written to hardcopy and that a site summary report has been written. The
User Followup Data File was also written upon at this point.
The program now proceeds to the next site. First it brings in the con-
centration data, then displays the analysis.
YEAR: 81 HATE: 0229
DATA SEARCH BEGUN FOR
SITE: BACKGROUND
SITE: URBAN CORE
SITE: BACKGROUND
CrtB SITE: BACKGROUND
2 UITH
iRESSIO
RIDG
•-UG/M3
.006+-
.334+-
214 + -
187+-
538 + -
839 + -
572+-
735 + -
088+-
053+-
147+-
884+-
596+-
4.025+-
HEASURED CONCENTRATION FINE/COARSE/TOTAL:
28.00000+- 3.000/ 37.00000+- 4.000/
rui
R-SQUARE:
aUUctUc
INTERCEPT
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
9 DIST OIL
10 RES GAS
11 COAL
13 ELARCFRN
14 FERRhNFR
16 CEMENT
17 PET FCC
19 SEC S04
fOJAL:
JUt KHU
.8287
m
-.
m
-,
12.
1.
1.
-.
1.
•
•
2.
4.
4.
27.
YEAR: 81
YEAR: 81
YEAR: 81
START HOUR:
E K= .120
.006
.346
.656
.093
7.325
.609
3.786
.733
.951
.114
.137
1.166
4.504
4.481
7.882
DATE: 0229 FC
DATE: 0229 FC
DATE: 0229 FRACTION: FINE
7 BACKGROUND: NO EFF VAR: NO
WtibnlcJJ LtnOt aUUrmtD —
R-SQUARE: .9241
lin/M'7 UTU — -
— _ mj/pj__ _ vir
.005+- .006
-14.654+- 34.854 7594.633
1.231+- 3.814 91.330
-.667+- .264 19.554
-26.433+- 39.874 52.511
2.261+- .522 1.823
20.923+- 12.055 25.512
17.660+- 61.095 6725.402
3.035+- 2.132 13.675
-.117+- .974 93.757
.186+- .493 22.554
9.390+- 3.855 30.223
6.337+- 24.170 55.903
7.995+- 30.701 78.138
27.147+- 20.851
65.00000+-
.000
84
-------
The user decides to stop the AUTOFIT sequence and return to command
mode.
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COHMAND
>yrop
ENTER COHMAND
The user is now in command mode and may enter any valid command. How-
ever, he decides that he would like to continue the AUTOFIT sequence after
all. He enters the RESUME command and the programs takes up the AUTOFIT
sequence where it left off.
>RESUME
URITTEN
CMB SITE: BACKGROUND YEAR: 81 DATE: 0229 FRACTION: COARSE
SAMPLE DURATION: 12 UITH START HOUR: 7 BACKGROUND: NO EFF VAR: NO
RIDGE REGRESSION UEIGHTED LEAST SQUARES
R-SQUARE: .8785 RID6E K= .480 R-SQUARE: .9570
SOURCE UG/M3 UG/H3 VIF
INTERCEPT .003+- .002 .003+- .003
1 SOIL 20.253+- 5.861 27.211+- 16.919 7.651
2 RD DUST 6.130+- 1.363 7.989+- 2.435 3.687
3 SEA SALT -.024+- .031 -.053+- .114 10.065
5 AUTO CAT 1.749+- 4.822 -9.926+- 14.544 8.157
6 AUTO .540+- .106 .820+- .099 1.130
11 COAL 1.397+- .840 2.655+- 1.708 4.580
12 KRAFT RB .053+- .538 -.156+- 2.073 10.701
16 CEMENT 6.954+- 2.694 9.610+- 3.238 1.852
TOTAL: 37.055+- 13.502 38.153+- 21.236
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
28.00000+- 3.000/ 37.00000+- 4.000/ 65.00000+- 5.000
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
>C
WRITTEN
URITTEN
AUTOFIT SEQUENCING FINISHED
ENTER COHMAND
85
-------
The program has noted that the AUTOFIT sequence is finished.
The user would now like to average the previous analyses across sites
(or dates if there were multiple dates). This is done with the PCOMP command.
Only site analyses which were followed by the double "WRITTEN" message are
included. Ordinarily, the results to be averaged are of a comparable
nature, e.g., averages could be computed for a set of sites in the urban
core for one or more dates. However, averaging the urban core and back-
ground sites illustrates the mechanics.
>PCOMP
OUTPUT UILL BO TO HARDCOPY.
DO YOU UftNT IT DISPLAYED AT YOUR TERMINAL INSTEAD?
.>YE3
CHB SITE
URBAN CORE
BACKGROUND
DATE
81/02/29
81/02/29
SOURCE
SOIL
SOIL
AVERAGE
(STD. DEV.)
FINE
-------
CHB SITE
URBAN CORE
BACKGROUND
CHB SITE
URBAN CORE
BACKGROUND
CHB SITE
URBAN CORE
BACKGROUND
CMB SITE
URBAN CORE
BACKGROUND
DATE SOURCE
81/02/29 SEA SALT
81/02/2? SEA SALT
AVERAGE
(STD. DEV.) +-
DATE SOURCE
81/02/29 AUTO CAT
81/02/29 AUTO CAT
AVERAGE
(STD. DEV.) +-
DATE SOURCE
81/02/29 AUTO
81/02/29 AUTO
AVERAGE
(STD. DEV.) +-
DATE SOURCE
81/02/29 DIST OIL
81/02/29 DIST OIL
AVERAGE
(STD. DEV.) +-
FINE
(UG/M3)
-.45
-.19
-.32
.18
FINE
(U6/M3)
7.38
12.54
9.96
3.65
FINE
(UG/M3)
4.63
1.89
3.26
1.94
FIXE
(UG/H3)
3.60
1.57
2.58
1.43
COARSE
(UG/H3)
.18
-.02
.08
+- .14
COARSE
(U6/N3)
10.34
1.75
6.04
+- 6.07
COARSE
(UB/M3)
.72
.54
.63
+- .13
COARSE
(UG/H3)
.00
.00
.00
+ - .00
TOTAL"
(UG/M3)
-.27
-.21
-.24
+- .04
TOTAL
(UG/M3)
17.72
14.29
16.00
+- 2.43
TOTAL
(U6/H3)
5.36
2.43
3.89
+- 2.07
TOTAL
(UG/M3)
3.60
1.57
2.58
+- 1.43
87
-------
CHB SITE
URBAN CORE
BACKGROUND
CHB SITE
URBAN CORE
BACKGROUND
CMB SITE
URBAN CORE
BACKGROUND
CHB SITE
URBAN CORE
BACKGROUND
DATE SOURCE
81/02/29 RES GAS
81/02/29 RES 6AS
AVERAGE
(STD. DEV.) -f-
DATE SOURCE
81/02/29 COAL
81/02/29 COAL
AVERAGE
(STD. DEV.) +-
DATE SOURCE
81/02/29 KRAFT RB
81/02/29 KRAFT RB
AVERAGE
(STD. DEV.) +-
DATE SOURCE
81/02/29 ELARCFRN
81/02/29 ELARCFRN
AVERAGE
(STD. DEV.) •*•-
FINE
(UG/M3)
1.02
-.73
.14
1.24
FINE
(UG/M3)
1.77
1 .09
1.43
.48
FINE
(UG/M3)
.00
.00
.00
.00
FINE
(UG/M3)
.08
-.05
.01
.09
COARSE
(UG/H3)
.00
.00
.00
+- .00
COARSE
(UG/H3)
4.53
1.40
2.96
+- 2.22
COARSE
(UO/H3)
2.23
.05
1.14
+- 1.54
COARSE
(U6/H3)
.00
.00
.00
+- .00
TOTAL
(UG/H3)
1 .02
-.73
.14
+- 1.24
TOTAL
(UG/H3)
6.30
2.49
4.39
+- 2.70
TOTAL
(U6/M3)
2.23
.05
1.14
+- 1.5-4
TOTAL
(UG/H3)
.08
-.05
.01
+- .09
88
-------
ChB SITE
URBAN CORE
BACKGROUND
CMB SITE
URBAN CORE
BACKGROUND
CHB SITE
URBAN CORE
BACKGROUND
CMB SITE
URBAN CORE
BACKGROUND
DATE SOURCE
81/02/29 FERRMNFR
81/02/29 FERRMNFR
AVERAGE
-------
The user would like to know what the current status is. That is, what
are the current site, size fraction, sources, and species. He enters the
PINFO command.
J-PINFO
***CURRENT STATUS***
CMB SITE: BACK6ROUND YEAR: 81 DATE: 0229
COARSE SIZE FRACTION
DURATION: 12 START HOUR: 7
HAS BACKGROUND BEEN SUBTRACTED: NO
FITTING SPECIES
.1 TOTAL
2 AL
4 BA
5 BR
6 CA
8 CL
9 CO
10 CR
11 CU
12 FE
14 K
15 m
17 P
18 PB
21 SI
23 SR
24 TI
26 IH
FITTING SOURCES
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTU
11 COAL
12 KRAFT RB
16 CEMENT
I£NTER COMMAND
90
-------
This shows that the last site and size fraction analyzed in the
AUTOFIT sequence is current.
The user decides to do an alternate analysis. He chooses ridge re-
gression as before, but he asks for the summary of all solutions. He also
sets the weight to 10.0, favoring the selection of a solution without sig-
nificant negative coefficients. Also, he chooses the effective variance
option.
.>CMfi
DO YOU WANT RIDGE SOLUTIONS?
>Y
ChB WILL SELECT THE BEST RIDGE SOLUTION
DO YOU WANT TO SEE A SUMMARY OF THE OTHERS?
>Y
IN THE SELECTION OF THE BEST RIDGE SOLUTION,
UHftT UEIBHT DO YOU WANT? (DEFAULTS .0)
WEIGHT IS 10.000000000
DO YOU UANT TO USE EFFECTIVE VARIANCES?
>Y
DO YOU UANT TO INCLUDE THE INTERCEPT IN THE MODEL?
>Y
Note that the user does not spell out "YES" but instead merely gives a
"Y." He could also use "N" for "NO." The first letter can be used for
most replies to the CMB program queries. It cannot be used for COMMANDS.
The CMB program now displays the summary of solutions.
91
-------
RtDGL K
.000
.005
.010
.015
.020
.025
.030
.035
.040
.045
.050
.060
.070
.080
.090
.100
.120
.140
.160
.180
.200
.240
.280
.320
.360
.400
.480
.560
.640
.720
.800
R-SQUARE
.eas/
.8836
.8837
.3838
.8838
.8839
.8839
.8839
.8839
.8338
.8838
.8837
.8835
.8832
.8829
.8326
.8817
.8806
.8793
.8778
.8762
.8725
.3682
.8635
.8584
.8530
.8414
.8291
.8164
.8033
.7900
AEROSOL
37.315
37.293
37.276
37.263
37.253
37.244
37.236
37.229
37.222
37.215
37.208
37.192
37.175
37.154
37.131
37.104
37.042
36.969
36.385
36.792
36.691
36.469
36.226
35.966
35.693
35.412
34.832
34.241
33.649
33.062
32.483
SE( AEROSOL)
16.01
15.91
15.83
15.75
15.67
15.60
15.52
15.45
15.38
15.31
15.24
15.11
14.99
14.87
14.75
14.64
14.42
14.22
14.03
13.85
13.68
13.37
13.08
12.82
12.58
12.35
11.93
11.56
11.22 •
10.90
10.61
tf NEG
3
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
9
2
2
2
2
o
2
2
2
2
1
UORST
-6.547
-5.666
-4.919
-4.267
-3.692
-3.181
-2 . 723
-2.310
-1.936
-1.596
-1.284
- . 735
-.274
-.271
-.267
-.262
-.252
-.240
-.228
-.216
-.204
-.181
-.159
-.139
-.120
-.103
-.072
-.046
-.023
-.013
-.012
ITERATIONS
3
T
2
1
i.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
9
2
2
9
2
2
2
RIDGE K= .180 IS THE SELECTED BEST SOLUTION
The R value is given for each value of the ridge K. The calculated
total aerosol mass and its standard error are also given. The program tries
to find a solution with this calculated value close to the measured value,
which is 35 Pg/m in this case. The number of negative coefficients is also
given along with the value of the negative coefficient of greatest magnitude.
Finally, the number of iterations in the effective variance procedure is
given.
92
-------
In conventional ridge regression, the R value is highest for the case
when k = 0. Due to certain required data transformations discussed in
^
Appendix D, however, R may oscillate in some cases as k increases if
effective variances are used. The R value is a measure of the agreement
between observed and predicted concentrations after the variance stabilizing
transformation.
Next the CMB program displays the chosen best solution.
CMS SITE: BACKGROUND YEAR: 61
SAMPLE DURATION: 12 UITH START HOUR:
RID6E REGRESSION
R-SQUARE: .8778 RIDGE K= .180
DATE: 0229 FRACTION: COARSE
7 BACKGROUND: NO EFF VAR: YES
WEIGHTED LEAST SQUARES
R-SQUARE: .8837
SUUKLt
INTERCEPT
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
11 COAL
12 KRAFT RB
16 CEMENT
TOTAL:
uu/nj
.003+-
15.349+-
7.963+-
-.031+-
2.453+-
.687+-
1.255+-
-.216+-
9.329+-
36.792+-
.002
6.147
2.190
.047
4.943
.206
.925
.794
3.380
13.852
uu/nj
.004+-
22.079+-
8.616+-
-.054+-
-6.547+-
.847+-
2.268+-
-.203+-
10.305+-
37.315+-
.003
19.624
3.282
.120
16.549
.236
2.178
1.980
5.092
16.005
9.394
1.881
4.668
8.510
1.261
2.815
4.853
2.009
(MEASURED CONCENTRATION FINE/COARSE/TOTAL:
28.00000+- 3.000/ 37.00000+- 4.000/ 65.00000+- 5.000
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COHMANH
ENTER COMHAND
The user would like to look at all of the solutions. He therefore
enters the PSOLN command.
93
-------
I'SOLN
THERE ARE 31 SOLUTIONS UITH AN INTERCEPT
AND 8 SOURCE COEFICIENTS AND STANDARD ERRORS
EACH FOR 52? NUMBERS TOTAL,
THEY CAN PRINT TO YOUR TERMINAL OR TO HARD COPr.
ARE YOU SURE YOU UANT TO LOOK AT THESE NUMBERS?
.>1
DO YOU UANT THE SOLUTIONS PRINTED ON THE HARD COPY
Y
SOLUTIONS WRITTEN TO HARD COPY
ENTER COMMAND
Recognizing the volume of data in the display, the user has elected to
send the results to the hardcopy. The results are included in Appendix C.
Note that the user is continuing to use shorthand replies.
The user would like to check some of the input data. He enters the
PMATRIX command.
>PMATRIX
MATRICES CAN GO TO HARD COPY
DO YOU UANT THEM DISPLAYED AT YOUR TERMINAL INSTEAD?
>Y
UHAT DO YOU UANT TO SEE?
SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS, OR CODES'?
.'-RECEPTORS
The user spelled out "RECEPTORS," but the abbreviation "R" could
have been used. The display is as follows:
-------
CURRENT MEASURED UG/M3 CONCENTRATIONS FOR
SITE.-BACKGROUND YEAR: 81 DATE: 0229
SPECIES
1 TOTAL
2 AL
3 AS
4 BA
5 BR
6 CA
7 CD
8 CL
9 CO
10 CR
11 CU
12 FE
13 HG
14 K
15 UN
16 NI
17 P
18 PB
19 304
20 SB
21 SI
22 SN
23 SR
24 TI
25 V
26 ZN
27 C
28 NA
29 N03
30 RB
31 SE
FINE
28.000+-
.070+-
.014+-
.146+-
.164+-
.128+-
.001+-
.010+-
.008+-
.010+-
.027+-
.120+-
.001+-
.186+-
.022+-
.003+-
.103+-
.894+-
14.780+-
.008+-
. 1 89+-
.024+-
.003+-
.008+-
.003+-
.038+-
.000+-
.000+-
.000+-
.000+-
.000+-
3.000
.055
.006
.018
.010
.008
.002
.007
.003
.005
.003
.008
.003
.010
.004
.002
.018
.047
.765
.003
.027
.003
.002
.OU
.009
.003
.001
.001
.001
.001
.001
COARSE
37.000+-
.739+-
.001+-
.086+-
,064+-
3.805+-
.001+-
.014+-
.001+-
.003+-
.004+-
.692+-
.009+-
.215+-
.018+-
.001+-
.020+-
.213+-
.852+-
.001+-
4.000+-
.003+-
.006+-
.097+-
.003+-
.016+-
.000+-
.000+-
.000+-
.000+-
.000+-
4.000
.171
.004
.016
.005
.444
.002
.010
.003
.005
.002
.044
.003
.032
.004
.002
.022
.017
.783
.003
.914
.003
.002
.017
.007
.003
.001
.001
.001
.001
.001
TOTAL
65.000+-
.809+-
.015+-
.232+-
.228+-
3.933+-
.002+-
.024+-
.009+-
.012+-
.031+-
.812+-
.010+-
.401+-
.040+-
.004+-
.123+-
1.107+-
15.632+-
.009+-
4.189+-
.027+-
.009+-
.105+-
.006+-
.054+-
.000+-
.000+-
.000+-
.000+-
.000+-
5.000
.180
.007
.024
.011
.444
.003
.012
.004
.007
.004
.045
.004
.034
.006
.003
.028
.050
1.095
.004
.914
.004
.003
.023
.011
.004
.001
.001
.001
.001
.001
The CMB program again asks the user what input data should be displayed.
This time the user selects the source signature matrix. However, he only
wishes to look at one of the sources, not the entire matrix:
95
-------
UHAT DO YOU UAHT TO SEE7
SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS, OR CODES?
OR ARE YOU DONE?
UHftT SIZE FRACTION? (FINE OR COARSE)
DO YOU UANT TO LOOK AT THE UHOLE
n IS 19 SOURCES BY 31 SPECIES
>N
UH1CH SOURCE DO YOU UANT?
GIVE SOURCE «
>2
SOURCE: RD DUST
MATRIX?
1 TOTAL
2 AL
3 AS
4 BA
5 BR
6 CA
7 CD
8 CL
9 CO
10 CR
11 CU
12 FE
13 HG
14 K
15 PIN
16 NI
i;7 P
18 PB
I 9 S04
20 SB
21 SI
22 S«
23 SR
24 ri
25 V
26 U
27 C
28 MA
2V N03
30 RB
3 I SE
1
.0000 t-
.0659 +-
.0000 +-
.0000 +-
.0001 +-
.0300 +-
.0000 •«•-
.0000 + -
.0000 + -
.0000 +-
.0000 +-
.0573 + -
.0000 +-
.0000 + -
.0010 +-
.0000 +-
.0000 +-
.0000 +-
.0000 +-
.0000 +-
.2840 +-
.0000 +-
.0000 +-
.0101 +-
.0003 +-
.0000 +-
.0489 +-
.0175 +-
.0002 +-
.0000 f-
.0000 +-
.0000
.0165
.0000
.0000
.0000
.0075
.0000
.0000
.0000
.0000
.0000
.0143
.0000
.0000
.0003
.0000
.0000
.0000
.0000
.0000
.0710
.0000
.0000
.0025
.0001
.0000
.0122
.0044
.0001
.0000
.0000
96
-------
UHICH SOURCE DO YOU UANT?
GIVE SOURCE tt
>
UHAT DO YOU UANT TO SEE?
SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS, OR CODES?
OR ARE YOU DONE?
>DONE
ENTER COMMAND
Since the user only wanted to look at the Road Dust source signature,
he gave a RETURN when asked for the next source. Furthermore, since he was
finished examining input data, he returned to command mode.
Now the user employes the INITIAL command to restore the original
selection of sources and species for the coarse size fraction. This returns
the coarse source and species lists to their condition at the end of the
initialization phase.
^INITIAL
SIZE IS COARSE
ENTER COMMAND
Now the user does a CMB analysis on the coarse fraction of the
"Background" site without using the ridge option. Note that the "ridge"
results and the least squares solution are identical.
DO YOU UANT RIDGE SOLUTIONS?
>N
DO YOU UANT TO USE EFFECTIVE VARIANCES?
>N
DO YOU UANT TO INCLUDE THE INTERCEPT IN THE MODEL?
>N
97
-------
CMB SITE: BACKGROUND YEAR: 81
SAMPLE DURATION: 12 UITH START HOUR:
RIDGE REGRESSION
R-SQUARE: .9553 RIDGE K= .000
SOURCE UG/M3
INTERCEPT
1 SOIL
2 RD DUST
3 SEA SALT
5 AUTO CAT
6 AUTO
7 JET AIR
9 DIST OIL
10 RES GAS
11 COAL
12 KRAFT RB
13 ELARCFRN
14 FERRMNFR
16 CEMENT
17 PET FCC
TOTAL:
.000+-
80.412+-
5.407+-
.863+-
54.731+-
.850+-
-17.418+-
-13.134+-
-50.557+-
-15.310+-
-10.988+-
2.736+-
-9.637+-
11.503+-
-2.459+-
37.000+-
.000
80.592
8.682
1.132
69.153
.117
53.437
18.948
67.151
21.876
13.717
3.373
11.824
9.054
14.132
21.769
MEASURED CONCENTRATION FINE/COARSE/TOTAL:
28.00000+- 3.000/ 37.00000+- 4.000/ 65.00000+-
PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT COMMAND
x:
ENTER COMKAND
DATE: 0229 FRACTION: COARSE
7 BACKGROUND: NO EFF VAR: NO
WEIGHTED LEAST SQUARES
R-SQUARE: .9553
-U6/M3 VIF
.000+- .000
80.412+- 80.592
8.682
1.132
69.153
.117
53.437
18.948
67.151
21.876
13.717
5.407+-
.863+-
54.731+-
.850+-
-17.418+-
-13.134+-
-50.557+-
-15.310+-
-10.988+-
2.736+-
-9.637+-
11.503+-
-2.459+-
37.000+-
110.800
27.837
691.696
420.117
1 .109
31.827
187.619
350.204
81.796
366.998
3.373 1338.700
11.824 1205.257
9.787
4.676
9.054
14.132
21.769
J.OOO
Since the user did not ask for ridge solutions, the "ridge" solution
on the left is identical to the least squares solution on the right. The
user also chose not to include the intercept in the model. Therefore, the
intercept is set exactly to zero. It is apparent from the negative values
with large magnitudes and the large standard errors that conventional
weighted least squares is inadequate to handle this complex set of sources.
It is clear from the variance inflation factors that this case is strongly
affected by multicollinearity. As is discussed in Section 2, reducing
the number of sources and using the ridge feature are two possible approaches
for obtaining a reasonable solution.
98
-------
Now the user has completed his session and enters the EXIT command to
end it,
>EXIT
The session is now over.
99
-------
REFERENCES
Belsey, D. A., Edwin, K., and Welsch, R. E. (1980) Regression Diagnostics:
Identifying Influential Data and Sources of Collinearity, pp. 92-93,
John Wiley & Sons, Inc., New York.
Cooper, J. A., Watson, J. G., and Huntzicker, J. J. (1979) Summary of the
Portland Aerosol Characterization Study (PACS), Annual Air Pollution
Control Association Meeting (Cincinnati, OH).
Dempster, A. P., Schatzoff, M., and Wermuth, N. (1977) A Simulation Study
of Alternatives to Ordinary Least Squares. J. Am. Statist. Ass.
72,, 77-106.
Draper, N. R. and Smith, H. (1981) Applied Regression Analysis,
John Wiley & Sons, Inc., New York.
Dzubay, T. G. and Hasan, H. (1982) Tentative Source Profiles for Use in
Hybrid Model. Environmental Sciences Research Laboratory, The United
States Environmental Protection Agency.
Farrar, D. E. and Glauber, R. R. (1967) Multicollinearity in Regression
Analysis: The Problem Revisited. Review Econ. and Statist. 4_9,
92-107.
Hoerl, A. E. and Kennard, R. W. (1970) Ridge Regression: Biased Estimation
for Nonorthogonal Problems. Technometrics 1J2, 55-67.
100
-------
Kowalczyk, G. S., Choquette, C. E., and Gordon, G. E. (1978) Chemical
Element Balances and Identification of Air Pollution Sources in
Washington, DC, Atmospheric Environment 12, 1143-1153.
Learner, E. E. (1973) Multicollinearity: A Bayesian Interpretation. Review
Econ. and Statist. 59. 371-380.
Marquardt, D. W. (1970) Generalized Inverses, Ridge Regression, Biased
Linear Estimation, and Nonlinear Estimation. Technometrics 12,
591-612.
Marquardt, D. W. and Snee, R. D. (1975) Ridge Regression in Practice. Am.
Statistn. 29, 3-20.
Morrison, D. F. (1967) Multivariate Statistical Methods, McGraw-Hill Book
Company, New York.
Obenchain, R. L. (1977) Classical F-Tests and Confidence Regions for Ridge
Regression. Technometrics 19, 429-439.
Swindel, B. F. (1974) Instability of Regression Coefficients Illustrated.
Am. Statistn. 28, 63-65.
Vinod, H. D. (1978) A Survey of Ridge Regression and Related Techniques for
Improvements over Ordinary Least Squares. Review Econ. and Statist.
60., 121-131.
Watson, J. G. (1979) Chemical Element Balance Receptor Model Methodology
for Assessing the Sources of Fine and Total Suspended Particulate
Matter in Portland, Oregon. Doctoral Disseration, Oregon Graduate
Center.
Wichern, D. W. and Churchill, G. A. (1978) A Comparison of Ridge Estimators.
Technometrics 20, 301-311.
101
-------
Williamson, H. J., Balfour, W. D., and Schmidt, C. E. (1982) Source
Apportionment through Weighted Ridge Regression. EPA 1982 Workshop
on Receptor Model Validation: Mathematical and Empirical (Rougemont,
NC, March, 1982).
102
-------
APPENDIX A
SOURCE PROGRAM LISTINGS
This appendix gives the FORTRAN source program listings of the CMS main
program and subprograms. These programs and their functions are as follows:
MAIN Program control and output
CMBR Perform CMS analysis
FETCH Search receptor concentration data file for selected site/date
INV1 Matrix inversion
COARS Compute coarse fraction from Total and Fine fractions
PERC Compute component percentages
103
-------
TABLE A.1 CMB MAIN PROGRAM
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
HI
42
43
44
45
46
47
48
C
c
C
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
°ROGRAM CMB MAIN
THIS PROGRAM WAS ADAPTED FROM THE CMBDEB
PROGRAM DEVELOPED FOR THE STATE OF OREGON DEE
BY THE OREGON GRADUATE CENTER FOR STUDY AND RESEARCH.
THE PROGRAM IS BASED ON THE WORK OF DR. J. WATSON'S CEB2 CODE
AS MODIFIED BY D. TORKELSON AS THE CMBDES CODE. THIS
ADAPTATION WAS DEVELOPED BY J. CORE AND W. HAMILTON
OF THE US EPA, OA2PS.RTP. N. CAROLINA,1981.
THE PROGRAM WAS ORIGIONALLY WRITTEN FOR THE PRIME 350
AND ADAPTED TO THE EPA'S SPERRY UNIVAC 1100 SYSTEM.
PROGRAM VARIABLES
AND NECESSARY INPUT/OUTPUT FILES ARE DESCRIBED BELOW.
IN 1982, THE CMB MAIN PROGRAM AND THE SUBPROGRAM
FETCH WERE MODIFIED AND ENHANCED. THE NEW
SUBPROGRAM CMBR REPLACED CEB2,
THEREBY ADDING A RIDGE REGRESSION OPTION.
THIS WORK WAS DONE UNDER CONTRACT TO US EPA
BY DR. H.J. WILLIAMSON AND DR. D.A. DUBOSE
RADIAN CORP. AUSTIN,TEXAS
EPA PROJECT OFFICERS.P. FREAS
REBUIRED INPUT FILES-
UNIT2, SONM - FILE OF SOURCE CODES AND
SOURCE MNEMONICS. THIS FILE DEFINES THE
SOURCES IN THE WORKING SOURCE MATRIX.
UNITS, PONM - FILE OF POLLUTANT CODES AND
POLLUTANT MNEMONICS. THIS FILE DEFINES THE
POLLUTANTS IN THE WORKING SOURCE MATRIX.
UNIT7, SORF - FILE OF FRACTIONAL POLLUTANT COMPOSITIONS
FOR FINE FRACTION SOURCES SIGNATURES
UNITS, SORC - FILE OF FRACTIONAL POLLUTANT COMPOSITIONS
FOR COARSE FRACTION SOURCES SIGNATURES
UNIT13, DATA - FILE OF FINE AND TOTAL FRACTION AMBIENT DATA
RECEPTOR CONCENTRATION DATA
104
-------
TABLE A.I CUB MAIN PROGRAM (CONTINUED)
49 C
50 C
51 C OUTPUT DISK FILES-
52 C
53 C
54 C UNIT11, CMBOUT - FILE OF FINE, COARSE AND TOTAL
55 C CMB RESULTS INCLUDING CHEMICAL AND FIT DATA.
56 C GENERAL PRINT FILE
57 C OPTIONAL OUTPUT FROM COMMANDS PMATRIX,PCOMP,PSOLN
58 C
59 C UNIT 9, SUMMRY - FILE OF SUMMARIZED FINE , COARSE AND TOTAL
60 C SOURCE COTRIBUTIONS
61 C SUMMARY PRINT FILE
62 C
63 C UNIT12, MAGSTO - FILE OF INPUT CARD DATA PLUS MASS AND
64 C POLLUTANT CONTRIBUTIONS FOR EACH SOURCE.
65 C USER FOLLOW UP FILE
66 C
67 C UNIT10.TEMP - FILE OF TEMPORARY INPUT DATA STORAGE
68 C
69 C INTERACTIVE TERMINAL FILES
70 C
71 C UNITS - TERMINAL READ
72 C
73 C UNIT6 - TERMINAL WRITE
74 c **********************************************************************
75 C
76 C
77 c ****************#*******VARIABLE HAMES**************************
78 C
79 C WORD HOLDER DATA ARRAYS
80 C ELMENT(IEL)-ELEMENT MNEMONICS
81 C SOUNAMC JSOURO-SOURCE MNEMONICS
82 C COMAND(ICMND)-COMMANDS
83 C COMND1-COMMAND READ FROM SCREEN
84 C SIZNAM-SAMPLE FRACTION DESCRIPTION
85 C YBACK- CURRENT BACKGROUND SUBTRACTION INDICATOR
86 C TEMPN- TEMPORARY NAME MATCHING HOLDER
87 C STNM-CURRENT SITE NAME
88 C BSTNM-CURRENT BACKGROUND SITE NAME
89 C
90 C
9 1 C SOURCE COMPOSITION ARRAYS
92 C A(IEL,JSOURC,ISIZ)-SOURCE COMPOSITIONS
93 C UACIEL,JSOURC,ISIZ)-UNCERTAINTIES OF A
94 C SCODEC JSOURO- SOURCE CODES
95 C PCODE(IEL)- POLLUTANT CODES
96 C
105
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
97 C
98 C COUNTERS AND LIMITS
99 C I-GENERAL COUNTER
100 C IEL-ELEMENT COUNTER
101 C KEL-FITTING ELEMENT COUNTER
102 C LEL-ELEMENT DELETE COUNTER
103 C NEL-TOTAL NUMBER OF SOURCE BUANTIFIED ELEMENTS
104 C MEL-NUMBER OF FITTING ELEMENTS
105 C JSOURC-SOURCE COUNTER
106 C KSOURC-FITTING SOURCE COUNTER
107 C LSOURC-SOURCE DELETE COUNTER
108 C NSOURC-TOTAL NUMBER OF SOURCES
109 C MSOURC-NUMBER OF FITTING SOURCES
110 C IFLAG-METHOD FLAG COUNTER(INVERSION ERROR)
111 C ICMND-COMMAND COUNTER
112 C NCMND-TOTAL NUMBER OF COMMANDS
113 C MELMAX-MAXIMUM NUMBER OF FITTING ELEMENTS
114 C MSOMAX-MAXIMUM NUMBER OF FITTING SOURCES
115 C MSOMIX-MSOMAX LESS ONE TO ALLOW FOR INTERCEPT
116 C
117 C
118 C
119 C
120 C SAMPLE IDENTIFICATION VARIABLES
121 C CMBID- UNIBUE DATA SET NUMBER OR NAME
122 C UDUR-SAMPLING DURATION IN HOURS
123 C USTART-SAMPLING START HOUR
124 C
125 C
126 C OPERATIONS VARIABLES
127 C UCNC(IEL,ISIZ)-URBAN OPERATIONS MEASURED CONCENTRATIONS
128 C UUCNC(IEL,ISIZ)-UNCERTAINTY OF UCNC
129 C CLCN(IEL,ISIZ)-CALCULATED CONCENTRATION
130 C UCLCN(IEL,ISIZ)-UNCERTAINTY OF CLCN
131 C UFMSS-URBAN FINE MASS LOADING
132 C UUFMSS-UNCERTAINTY OF UFMSS
133 C UCMSS- COARSE MASS
134 C UUCMSS - UNCERTAINTY
135 C UTMSS- TOTAL MASS
136 C UUTMSS- UNCERTAINTY
137 C MFLAGdEL,3)-MISSING DATA FLAG FOR EACH ELEMENT AND SIZE
138 C
1 39 C
140 C BACKGROUND SUBTRACTION VARIABLES
141 C -SAME AS URBAN PREFACED WITH A 'B'
142 C
143 C
144 C
106
-------
TABLE A.I CMB MAIN PROGRAM (CONTINUED)
145 C TEMPORARY OUTPUT VARIABLES
146 C PCNT-PERCENT
147 C UPCNT-UNCERTAINTY PERCENT
148 C FPCNT-FINE PERCENT
149 C UFPCNT-UNCERTAINTY
150 C CPCNT-COARSE PERCENT
151 C UCPCNT-UNCERT
152 C TPCNT-TOTAL PERCENT
153 C UTPCNT-UNCERTAINTY
154 C X-GENERAL
155 C UX-GENERAL
156 C TX-GENERAL TOTALS
157 C UTX-UNCERTAINTY
158 C RATIO-RATIO OF CALCULATED TO MEASURED
159 C URATIO-UNCERTAINTY OF RATIO
160 C
161 C
162 C
163 C FIT INITIALIZATION ARRAYS
164 C INITELCIED-INITIAL FITTING ELEMENTS
165 C INITSOCJSOURO-INITIAL FITTING SOURCES
166 C IMEL-NUMBER OF INITIAL FITTING ELEMENTS
167 C IMSO-NUMBER OF INITIAL FITTING SOUCES
168 C
169 C MISC.
170 C ELFT(IEL,ISIZ)-ELEMENT FITTING FLAG
171 C ELHOLD(KEL,ISIZ)-ELEMENT FITTING PLACE HOLDER
172 C SOUFTCJSOURC,ISIZ)-SOURCE FITTING FLAG
173 C SOHOLD(KSOURC,ISIZ)-SOURCE FITTING PLACE HOLDER
174 C IRTN-GO TO STATEMENT VARIABLE
175 C JRTN-GO TO STATEMENT VARIABLE
176 C UNITX-FORTRAN FILE UNIT NUMBER
177 C X-GENERAL REAL NUMBER
178 C ICNTRL-AUTOFIT/SELECT CONTROL CODE FOR FETCH
179 C SEL(ISIZ)-ELEMENTAL SOURCE CONTRIBUTIONS
180 C USEL(ISIZ)-UNCERTAINTY
181 C ISCTR-PRINTED SOURCE COUNTER
182 C ISHOLD(ISCTR)-ARRAY POSITIONS OF PRINTED SOURCES
183 C IBACK-CURRENT BACKGROUND SUBTRACTION INDICATOR
184 C DUM(20)-SCRATCH VECTOR
185 C SAVE-AUTOFIT TITLE LINE HOLDER
186 C
187 C
188 C BOOKKEEPING
189 C
190 C ANS-ANSWER TO PROGRAM BUERIES
191 C CHK-ANSWER TO PROGRAM QUERIES
192 C IBACK-BACKGROUND SUBTRACTION STATUS FLAG
107
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
193 C IWIDE-TERMINAL LINE WIDTH FLAG
194 C LASTID-HOLDER FOR CMBID IN PCOMP
195 C LOOP -CYCLING STATUS FLAG IN PMATRIX
196 C MANY-LINE WIDTH CONTROL INDEX
197 C MORE-LINE WIDTH CONTROL INDEX
198 C NOSKIP-CMB PARAMETER CONTROL FLAG CONTROLLED BY AUTOFIT
199 C OLDID-HOLDER FOR CMBID TO PREVENT TOTALING FINE AND COARSE
200 C OF DIFFERENT SITES IN WRITE
201 C PNAM-ARRAY FOR CMBID'S IN PCOMP
202 C SITE-SITE CODE
203 C YEAR-YEAR : COMPONENTS OF CMBID
204 C DATE-MONTH AND DAY :
205 C SPACE-DUMMY CHARACTER FOR READ
206 C TEMPID-TEMPORARY HOLDER FOR CMBID CHECK
207 C
208 C
209 C
210 C CMBR SUBPROGRAM ARGUMENTS AND RELATED VARIABLES
211 C
212 C AP-CMBR WORK SPACE
213 C C-RECEPTOR CONCENTRATION VECTOR
214 C CWT-WEIGHT FOR SELECTING RIDGE SOLUTIONS
215 C CWT10-TEMPORARY HOLDER FOR CWT VALUE
216 C EFVAR-EFFECTIVE VARIANCE FLAG
217 C SELK-LEAST SQUARES SOLUTION
218 C F-SOURCE SIGNATURE MATRIX
219 C IA-CMBR WORK SPACE
220 C INTCEP-INTERCEPT OPTION FLAG
221 C IS-CMBR WORK SPACE
222 C ITM-CMBR WORK SPACE
.223 C KINDEX-INDEX OF SELECTED BEST RIDGE SOLUTION
224 C R2LS-R-S2UARE FOR LEAST SQUARES SOLUTION
225 C RIDGE -RIDGE OPTION FLAG
226 C S-CMBR WORK SPACE
227 C SAINT-SELECTED RIDGE SOLUTION INTERCEPT
228 C SAINTO-LEAST SQUARES SOLUTION VECTOR
229 C SAS-RIDGE SOLUTION MATRIX
230 C SAA-RIDGE SOLUTION STANDARD ERROR MATRIX
231 C SET-STANDARD ERROR OF CALCULATED TOTAL AEROSOL
232 C SETLS-STANDARD ERROR OF CALCULATED TOTAL AEROSOL(LEAST SQUARES)
233 C SF-STANDARD ERROR OF SOURCE SIGNATURE MATRIX
234 C SOLD-CMBR WORK SPACE
235 C SR2-R-SQUARE FOR SELECTED RIDGE SOLUTION
236 C SSET-STANDARD ERROR OF CALCULATED TOTAL AEROSOL
237 C SX-CMBR WORK SPACE
238 C USAINT-UNCERTAIHTY OF SAINT (RIDGE SOLN INTERCEPT)
239 C USAINO-UNCERTAINTY OF SAINTO (LEAST SQUARES INTERCEPT)
240 C VELK-STANDARD ERROR OF LEAST SQUARES SOLUTION
108
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
241 C TM-WORK VECTOR FOR CMBR
242 C TOTALA-TOTAL AEROSOL MASS
243 C V-STANDARD ERRORS OF RECEPTOR CONCENTRATION VECTOR
244 C VIF-VARIANCE INFLATION FACTOR
245 C VIFELK-VIF ORGANIZED BY SOURCE INDECIES
246 C VK-VECTOR OF K-VALUES FOR RIDGE REGRESSION
247 C XMEAN-USED BY CMBR FOR SOURCE MEANS
248 C XP-USED BY CMBR FOR X MATRIX
249 C XPX-USED BY CMBR FOR X'X MATRIX
250 C XPXK-USED BY CMBR FOR X'X MATRIX(INVERSE)
251 C XPY-USED BY CMBR FOR X'Y VECTOR
252 C YP-USED BY CMBR FOR Y VECTOR
253 C
254 C
255 C CHARACTER TEST CONSTANTS
256 C BLANK,IBLANK-BLANK '
257 C ISTAR-PRINT FLAGS ' ' '*'
258 C BUOTA-'A' QUOTB-'B1 gUOTC-'C' (COARSE,CODES)
259 C SUOTD-'D' (DONE)
260 C eUOTOL-'TOTAL' BUOTR-'R' (RECEPTOR)
261 C eUOTS-'S1 (SOURCE,STOP) STAR-ASTERISK '*'
262 C YESY-'Y' (YES)
263 C
264 C ***********************************************************:*:*;':***;*:*:'
265 C
266 C
267 C
268 C
269 C SPECIFICATION AND DIMENSION
270 C
271 REAL*8 COMAND(20),SIZNAM(3),SOUNAM(35),ELMENT(35),TEMPN,
272 *COMND1,SUOTOL
273 CHARACTER VARIABLES
274 INTEGER*4 CHK,ELFT(35,2),SOUFT(35,2),ANS,
275 *YBACK(2),TYPE,YESY,
276 * MFLAG(35,3) ,MFLAGB(35,3),DUM(20),ISTAR(3) ,
277 * STAR,eUOTC,2UOTA,eUOTB,£UOTS,SPACE,CMBID(S),BACKID(5),
278 * OLDID(5),fiUOTR,BUOTD,TEMPID(6),LASTID(6),
279 *SAVE(5), FLAG(8,3),YEAR,BYEAR,SITE(3),BSITE(3),DATE,BDATE
280 C
281 C NUMERIC VARIABLE SPECS
282 INTEGER*4 ELHOLD(35,2),SOHOLD(35,2 ) ,
283 *INITEL(21),INITSO(16),IMEL,IMSO,
284 *MEL1(2) .MSORCK2) , USTART,UDUR,
285 *UNIT,NEL,NSOURC,MELMAX,MSOMAX,BSTART,BDUR,
286 *IRTN,JRTN.IEL,JSOURC,NCMND,I,KEL,LEL, ICMND,UNIT6,UNIT 10,UNIT 13
287 *UNIT12,MEL,MSOURC,KSOURC,LSOURC,
288 *EFVAR,SIZ,UNITS,UNITX,UNIT9,
109
-------
TABLE A.I CMB MAIN PROGRAM (CONTINUED)
289 *SCODE(35),PCODE(35),ISHOLD(35),UNIT!1,
290 *TEMPC, IFLAG(3),RIDGE
29 1 C
292 C
293 REAL*4 A(35,35,2),UA(35,35,2),SEL(2),USEL(2 ) ,
294 *USORC(35,3),UUSORCC35,3),UCNC(35,3>,UUCNC(35,3),
295 *CLCN(35,3),UCLCN(35,3),UFMSS,UUFMSS,RATIO,URATIO,
296 *TX,UTX,X,UX,PCNT,UPCNT,UTMSS,UUTMSS,UCMSS,UUCMSS,
297 *BFMSS,UBFMSS,UBTMSS,BTMSS,BCMSS,UBCMSS,BCNC(35,3),UBCNCC35,3),
298 * SELKC35),UELK(35),VIFELK(35),SAINT(3),SETC3),USAINT(3)
299 C
300 C
301 C CHEMICAL ELEMENT BALANCE ARRAYS
302 C
303 C STORAGE FOR CMB SUBROUTINE (17 SOURCES, 21 SPECIES)
304 C THE INTERCEPT COUNTS AS A SOURCE
305 REAL F(21 , 17) ,C(21),V(21 ),SF(21 , 17) , VK(31), W(21),
306 * YP(21),XP(21,17),XMEAN(17), XPX(17,17),
307 * XPXK(17, 17),XPY(17),SX(17),S(17),TM(21),AP(17),SOLD(17) ,
308 * VIF(17),SAS( 17,31),SAA( 17,31 ) ,SSET(2)
309 INTEGER IS(17),IA(17),ITM(21),PNAM(17,6)
310 REAL*8 BLANK
311 C
312 C
313 C EQUIVALENCE STATEMENT
314 EBUIVALENCE (IS,S),(IA,AP),(ITM,TM)
315 E2UIVALENCE (IBLANK,BLANK),(STAR,ISTAR(2)),
316 * (XPXK,PNAM),
317 * (CMBID,SITE),(CMBID(U),YEAR),(CMBID(5),DATE),
318 * (BACKID,BSITE),(BACKID(4),BYEAR),(BACKID(5),BDATE)
319 C
320 C
321 C COMMON
322 COMMON /F/ PCODE,SCODE,NEL
323 C
324 C
325 C DATA STATEMENTS
326 DATA YBACK/'YES ','NO '/,YESYX•Y*/
327 C
328 DATA BLANK/' '/,6UOTC,QUOTA,2UOTB,2UOTS/
329 * 'C','A','B','S'/,eUOTOL/'TOTAL'/,
330 * 2UOTR,2UOTD/'R','D'/
331 DATA COMAND/
332 *'HELP ' ,
333 *'AE ','DE ','AS ',•DS ','CMB
334 *'PSOLN ','PINFO ','PDATA ','WRITE
335 ^'INITIAL ','PMATRIX ','PCOMP ','RESUME ',
336 *'AUTOFIT '.'SELECT ','BACKOUT ',
1 10
-------
TABLE A.I CMB MAIN PROGRAM
(CONTINUED)
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
38 1
382
383
384
C
C
C
C
c :
C S;
C 1
C 1
c
c :
C D
C B
C
C
C
c
c c
1 10
1 15
120
*'BACKIN
EXIT
'SIZE
DATA VK /O.0,0.005,0.01,0.015,0.020,0.025,
* 0. 030, 0.035,0. QUO, 0 .045, 0.050, 0.060,0.070,0.080,0.090,
* 0.100, 0.12,0.14,0.16,0.18,0.20, 0.24,0.28,0.32,0.36,
* 0.40, 0.48,0.56,0.64,0.72,0.80 /
DATA SIZNAM/" FINE
COARSE
TOTAL
DATA
*
*
FLAG/'
* — — _
'RIX '
'VERG'
t t
'IS S'
'ENT
,' MAT'
, ' CON'
«,'** '/SNOSKIP/O/
PARAMETERS
ELEMENTS
OF FITTING SOURCES
1 CEB'
' ',' NON'
DATA ISTAR/' ','*
INITIALIZATION OF STARTING
OUTPUT UNIT NUMBERS
UNIT5=5
UNIT6=6
UNIT9=9
UNIT10=10
UNIT1 1 = 11
UNIT12=12
UNIT13=13
MAXIMUM NUMBER OF FITTING
MELMAX=2 1
MAXIMUM NUMBER
MSOMAX=17
MSOMIX=MSOMAX-1
SET RETURN STATEMENT
IRTN=1
NUMBER OF COMMANDS
NCMND=20
SEARCH CONTROL SET TO SELECT
ICNTRL=4
BACKGROUND SUBTRACT SET TO 'NONE'
IBACK=2
- CREATE WORKING SOURCE MATRICES -
* f
CODE SOURCE NAMES AND POSITIONS FROM SONM FILE
NSOURC=0
DO 140 1=1,35
SOUNAM(I)=BLANK
SCODE(I)=0000
READC 2, 120,ERR=130,END=150)TEMPC,TEMPN
FORMATCI2,2X,AS)
SOUNAM(I)=TEMPN
INGU*,'LAR '
','SOLU','TION'
MODE
1 1 1
-------
TABLE A.1 CMB MATH PROGRAM
(CONTINUED)
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
41 1
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
C ERI
130
135
140
c
>
C
10!
]>
I
* •
* •
(
(
C CODE
150
160
170
C DO
175
c ER:
180
185
190
200
C IN
201
210
C CO
215
1
I
]
]
1
Ni
i
1
RO
*
*
IT
DE
CARD DID NOT MATCH
WAS IGNORED. CHECK
PROPER FORMAT'/
THIS AT UPCOMING '/
SCODE(I)=TEMPC
NSOURC=NSOURC+1
GO TO 140
MESSAGE- FORMAT MISMATCH
WRITECUNIT6,135)
FORMATMH , ' AN INPUT
'IN FILE SONM AND
*1X,'VERIFY POINT')
GO TO 115
CONTINUE
POLLUTANT NAMES AND POSITIONS FROM PONM FILE
NEL=0
KINDEX=0
DO 190 1=1,35
ELMENT(I)=BLANK
PCODE(I)=000
READC 3, 170,ERR=180,END=200)TEMPC,TEMPN
FORMAT(I2,2X,A8)
OT CODE IF POLLUTANT IS MASS CONCENTRATION
IFCTEMPC.NE.01)GO TO 175
TEMPN=2UOTOL
KINDEX=1
ELMENT(I)=TEMPN
PCODECI)=TEMPC
NEL =NEL+1
GO TO 190
>R MESSAGE- FORMAT MISMATCH
WRITECUNIT6,185)
FORMATC1H ,'AN INPUT
*1X,'IN FILE PONM AND
*1X,'VERIFY POINT')
GO TO 160
CONTINUE
IFCKINDEX-EB.1)GO TO
PCODE(I)=1
ELMENT(I)=2UOTOL
NEL=NEL+1
IALIZE SOURCE MATRICES
DO 210 ISIZ=1,2
DO 210 I=1,NSOURC
DO 210 J=1,NEL
A(J,I,ISIZ)=O.OOOE-00
UA(J,I,ISIZ)=0.001E-02
CONTINUE
I MATRICES
DO 270 ISIZ=1,2
IUNIT=ISIZ+6
READCIUNIT,220,ERR=232,END=270)ISCODE,IPCODE,X,UX
CARD DID NOT MATCH PROPER FORMAT'/
WAS IGNORED. CHECK THIS AT UPCOMING '/
201
1 12
-------
TABLE ' CMB MAIN PROGRAM (CONTINUED)
433 220 FORMATCI2,2X,I2,2X,F8.6,2X,F8.6)
434 IFCIPCODE.E2.1)WRITE(UNIT6,221)IUNIT
435 221 FORMATC' **WARNING** POLLUTANT CODE 1 IN DATA ON UNIT ',
436 * 12,/' CODE 1 IS RESERVED FOR TOTAL AEROSOL')
437 C MATCH SCODES
438 DO 230 J=1,NSOURC
439 IFCISCODE.E2.SCODECJ))GO TO 240
440 230 CONTINUE
441 C ERROR MESSAGE - NO SCODE MATCH
442 232 IFCISIZ.E2. 1 ) WRITE( UNIT6 , 234)
443 234 FORMATC1H , 'FORMAT MISMATCH OR SOURCE CODE NOT IDENTIFIED FOR '/
444 *1X,'AN INPUT CARD IN SORF FILE . CARD IGNORED.')
445 WRITECUNIT6,235)ISCODE
446 235 FORMATCIX,'SOURCE CODE :',I3)
447 IFdSIZ.ES. 2)URITE(UNIT6, 236)
448 236 FORMATC1H ,'FORMAT MISMATCH OR SOURCE CODE NOT IDENTIFIED FOR '/
449 *1X,'AN INPUT CARD IN SORC FILE . CARD IGNORED.')
450 GO TO 215
451 C MATCH PCODES
452 240 DO 250 K=1,NEL
453 IF(IPCODE.Ee.PCODECK))GO TO 260
454 250 CONTINUE
455 C ERROR MESSAGE - NO MATCH
456 IF (ISIZ.E2.1)WRITE(UNIT6,252)SOUNAM(J)
457 252 FORMATC1H ,'AN UNIDENTIFIED POLLUTANT CARD WAS FOUND IN'/
458 *1X,'FILE SORF , SOURCE-',A5)
459 IFCISIZ.E2.2)WRITECUNIT6,254)SOUNAMCJ)
460 254 FORMATC1H ,'AN UNIDENTIFIED POLLUTANT CARD WAS FOUND IN'/
461 *1X,'FILE SORC , SOURCE-',A5)
462 GO TO 215
463 260 ACK,J,ISIZ)=X
464 UACK,J,ISIZ)=UX
465 GO TO 215
466 270 CONTINUE
467 C ADD TOTAL AEROSOL ELEMENT
468 DO 272 K=1,NEL
469 IFCPCODECK).E2.1) GO TO 273
470 272 CONTINUE
471 273 DO 274 ISIZ=1,2
472 DO 274 J=1,NSOURC
473 ACK,J,ISIZ) = 1 .0
474 UACK,J,ISIZ)=0.000001
475 274 CONTINUE
476 C WRITE DATA TO SCREEN FOR INSPECTION
477 278 WRITECUNIT6,300)
478 300 FORMATCIX,'WOULD YOU LIKE TO LOOK AT THE SOURCE AND SPECIES '
479 *'LISTS?'
480 * ,/,2X,'TYPE YES OR NO.')
1 13
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
481 301 READ(UNIT5,28001,ERR=278)ANS
482 302 FORMATCA3)
483 IF(ANS .NE. YESY )GO TO 335
484 GO TO 304
485 304 WRITE(UNIT6,310)
486 310 FORMATMH ,'SOURCE SIGNATURE MATRICES ARE CODED AS FOLLOWS1,/
487 *1X,'SOURCES:',/,
488 *1X,'SOURCE # SOURCE NAME',/)
489 DO 315 1=1,NSOURC
490 WRITECUNIT6,312)1,SOUNAM(I)
491 312 FORMAT(3X,I2, 1 IX,A8)
492 315 CONTINUE
493 WRITE(UNIT6,320)
494 320 FORMATMH ,/, 1X ,'SPECIES ••',/,
495 *1X,'SPECIES # SPECIES NAME ')
496 DO 330 1=1,NEL
497 WRITECUNIT6,312)1,ELMENT(I)
498 330 CONTINUE
499 C
500 C
501 C
502 C
503 C - SET UP INITIAL AUTOFIT POLLUTANTS AND SOURCES -
504 C
505 C
506 C
507 C INITIALIZE FIT FLAGS
508 335 DO 340 IEL=1,NEL
509 ELHOLDdEL, 1 )=0
510 INITEL(IEL)=0
511 340 ELFTCIEL,1)=IBLANK
512 DO 350 JSOURC=1,NSOURC
513 SOHOLDCJSOURC,1)=0
514 INITSO(JSOURC)=0
515 350 SOUFTCJSOURC,1)=IBLANK
516 C INPUT INITIAL AUTOFIT SOURCES AND ELEMENTS
517 360 WRITECUNIT6,365)
518 365 FORMATUH ,'PLEASE INPUT INITIAL FITTING INFORMATION')
519 IMSO=0
520 DO 370 I=1,NSOURC
521 366 WRITECUNIT6,367)1
522 367 FORMATC1HO,'INITIAL SOURCE ',12,': XX')
523 READ(UNIT5,2420,ERR=366)INITSOCI),SPACE
524 IF (SPACE.EB.IBLANK)INITSO(I)=INITSO(I)/10
525 IF(INITSOd) . E2 . 0)GO TO 380
526 IF(INITSOd) .E2.-1 )GO TO 372
527 IMSO=IMSO+1
528 370 CONTINUE
1 14
-------
TABLE A.I CMB MAIN PROGRAM (CONTINUED)
529 GO TO 380
530 372 IMSO--MINO(NSOURC,MSOMIX)
531 DO 373 1=1,NSOURC
532 373 INITSO(I)=I
533 380 WRITE(UNIT6,381)
534 381 FORMAT(' INITIAL SOURCES')
535 DO 390 I=1,IMSO
536 J=INITSO(I)
537 WRITE(UNIT6,425)J,SOUNAM(J)
538 390 CONTINUE
539 IMEL=0
540 WRITECUNIT6,400)
541 400 FORMATC' ',/, 1 X,'PLEASE INPUT SPECIES')
542 DO 410 1=1,NEL
543 403 WRITECUNIT6,404)1
544 404 FORMATC 1H , 'INITIAL SPECIES ',I2,'-' XX')
545 READ(UNIT5,2420,ERR=403)INITEL(I),SPACE
546 IF(SPACE.ES.IBLANK)INITEL(I)=INITEL(I)/10
547 IF(INITEL(I).E2.0)GO TO 420
548 IF(INITEL(I).E6.-1)GO TO 411
549 IMEL=IMEL+1
550 410 CONTINUE
551 GO TO 420
552 411 IMEL=MINO(NEL,MELMAX)
553 DO 412 I=1,IMEL
554 412 INITEL(I)=I
555 420 WRITE(UNIT6,421)
556 421 FORMAT(' INITIAL SPECIES')
557 DO 430 I=1,IMEL
558 J=INITEL(I)
559 WRITE(UNIT6,425)J,ELMENT(J)
560 425 FORMAT(f ',3X,12,3X,A8)
561 430 CONTINUE
562 C OPERATOR DECISION- CORRECT INPUT OR CONTINUE
563 WRITE(UNIT6,435)
564 435 FORMAT(' '/,1X,'ARE INITIAL SOURCES AND SPECIES CORRECT?')
565 READ(UNIT5,28001)CHK
566 IF(CHK.E2.YESY)GO TO 441
567 WRITE(UNIT6,442)
568 442 FORMAT(' USE COMMANDS AE,DE,AS,DS FOR CHANGES.'/
569 * ' OR, DO YOU WANT A FRESH START?')
570 READ(UNIT5,28001)ANS
571 IF(ANS.E2.YESY)GO TO 360
572 441 IRTN=0
573 SIZ=1
574 443 GO TO 4401
575 444 IF(SIZ.GE.2)GO TO 445
:;76 SIZ = 2
1 15
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
61 1
612
613
614
615
616
617
6 18
619
620
621
622
623
624
445
446
C
C
450
C
C
C
C
C
C
C
C
GO TO 443
IRTN=1
WRITE(UNIT6,446)
FORMAT(' THE "HELP" COMMAND LISTS COMMANDS')
SIZ=1
WRITE(UNIT6,31 10) SIZNAM(SIZ)
REWIND UNIT 10
GO TO (500,2860,4120,4130,4132,4150,4250,4260,
- START OF GENERAL INTERACTIVE NETWORK -
C COMMAND SEBUENCE
490
495
500
510
520
550
WRITE(UNIT6,495)
FORMATMH ,' COMMAND NOT EXECUTED. TRY AGAIN')
WRITE(UNIT6,510)
FORMATMH , 'ENTER COMMAND')
READ ( UNITS , 520 , ERR=490 ) COMND 1
FORMAT(AS)
DO 550 ICMND=1 , NCMND
IF(COMND1 .EQ.COMAND(ICMND) )GO TO 600
CONTINUE
GO TO 490
C BRANCHING
600
C
C
C
C
1 100
1 1 10
GO T0(
*1100, 2400, 2500, 2600, 2700, 2800,
*42000, 3000, 3200, 3400, 4400,
*43000, 52000, 4000,
*4099, 4200, 4600, 47 00, 5000, 3100
*) ,ICMND
-HELP. LIST AND DEFINE AVAILABLE COMMANDS
WRITE(UNIT6, 1110)
FORMAT( 1H ,
*'HELP-LISTS THESE COMMANDS ',/ 1 H ,
*' ' ,/1H ,
*• DATR ACCESS AND SEBUENCING ',/1H ,
*' AUTOFIT-SE6UENCES AUTOMATICALLY TO NEXT DATA
*'RESUME-RESUME AUTOFIT',/1H ,
*'SELECT-SELECT DATA SET FOR CMB',/1H ,
4270,4280),IRTN
SET',/1H
1 16
-------
TABLE A.1 CUB MAIN PROGRAM (CONTINUED)
625 *'SIZE-CHANGE SIZE FRACTION1/' ',
626 *'EXIT-CLOSE FILES AND LEAVE',/1H ,
627 *' ',/1H ,
628 *' CMB OPERATIONS ',/1H ,
629 *'AE-ADD A SPECIES TO THE FIT',/1H ,
630 *'DE-DELETE A SPECIES FROM THE FIT',/1H ,
631 *'AS-ADD A SOURCE TO THE FIT',/1H ,
632 *'DS-DELETE A SOURCE FROM THE FIT',/1H ,
633 *'CMB-PERFORM CMB •' OPTIONS ARE RIDGE £ EFFECTIVE VARIANCE'/' ')
634 WRITE(UNIT6, 1 120)
635 1120 FORMATC1H ,
636 *• ',/1H ,
637 *' SCREEN DISPLAY ',/1H ,
638 *'PINFO-PRINT CURRENT STATUS ON SCREEN',/1H ,
639 *'PDATA-PRINT CURRENT CMB RESULTS ON SCREEN',/1H ,
640 *'PMATRIX-PRINT SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS,'/
641 * ' OR SOURCE AND SPECIES CODE LISTS'/' ',
642 *'PSOLN-PRINT ALL RIDGE SOLUTIONS',/' ',
643 *'PCOMP-PRINT COMPUTED AVERAGES OF CMB SERIES',/' ',
644 *' ',/1H ,
645 *' DATA STORAGE ',/1H ,
646 *'WRITE-WRITE PRESENT CMB RESULTS TO FULL PRINTOUT,',/1H ,
647 *' SUMMARY,AND USER FOLLOWUP STORAGE FILES')
648 WRITE(UNIT6,1140)
649 1140 FORMATMH ,
650 *' ',/lH ,
651 *' BACKGROUND SITE OPERATIONS ',/1H ,
652 *'BACKOUT-SELECT AND SUBTRACT A BACKGROUND DATA SET',/1H ,
653 *'BACKIN-ELIMINATE CURRENT BACKGROUND SUBTRACTION')
654 WRITECUNIT6,1150)
655 1150 FORMATC//1 SELECT SHOULD NORMALLY BE THE FIRST COMMAND')
656 GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
657 C
658 C
659 C
660 C
661 C -AE. ADD AN SPECIES TO THE FIT
662 2400 JRTN=1
663 WRITECUNIT6,3110)SIZNAMCSIZ)
664 GO TO 2410
665 2405 WRITECUNIT6,2407)
666 2407 FORMATC1H ,'IMPROPER CODE, TRY AGAIN')
667 2410 WRITE(UNIT6,2415)
668 2415 FORMATC1H ,'INPUT CODE OF ADDED SPECIES')
669 READCUNIT5,2420,ERR=2405)IEL,SPACE
670 2417 IFCIEL.E2.0)GO TO C500,2860,4120,4130,4132,4150,4250,4260,
671 *4270,4280),IRTN
672 2420 FORMATC I2,T2,A1)
1 17
-------
TABLE R.I CMB MAIN PROGRAM
(CONTINUED)
673
67't
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
7 10
71 1
712
713
714
715
716
7 17
718
719
720
C Cfr
2425
2430
2480
2490
C
C
C
C
C
2500
2505
2510
2515
2517
C CH
2523
2525
2530
2540
2560
SPECIES
IF (SPACE. EC.IBLANK)IEL=IEL/10
IFdEL.LT. 1 .OR. IEL. GT.NEDGO TO 2405
'CK TO SEE IF THIS IS ALREADY A FITTING
IF(ELFT(IEL,SIZ) . E2 . IBLANK ) GO TO 2430
WRITE(UNIT6,2425)ELMENT(IEL)
FORMATMH ,A8,' IS ALREADY A FITTING SPECIES')
GO TO (2410,2510,26 10 , 2710,3402) ,JRTN
IF(MEL1 (SIZ) .E2.MELMAX)GO TO 2480
MEL1 (SIZ)=MEL1 (SIZ)+1
MEL=MEL1 (SIZ)
ELFT (IEL, SIZ )= STAR
ELHOLD(MEL,SIZ)=IEL
GO TO (2410,2510,2610,2710,3402) ,JRTN
WRITE (UNIT 6, 2490 )MELM AX, ELMENT(IEL)
FORMATMH , 'MAXIMUM OF ',12,' FITTING SPECIES EXCEEDED. ',
:1X,A4,' NOT ADDED TO FIT')
GO TO (500,2860, 4120,4130, 4132,4150,4250,4260,4270,4280) ,IRTH
-DE.
JRTN=2
WRITE(UNIT6
GO TO 2510
WRITE(UNIT6
WRITE(UNIT6
FORMATdH ,
READ(UNIT5,
DELETE AN SPECIES FROM THE FIT
,31lO)SIZNAM(SIZ)
,2407)
,2515)
'INPUT CODE OF DELETED SPECIES'
2420,ERR=2505)IEL,SPACE
IF(SPACE.Eg.IBLANK)IEL=IEL/10
IFdEL.EE.0)GO TO (500,2860,4120,4130,4132,4150,4250,4260
*4270,4280),IRTN
IFdEL.LT. 1 .OR.IEL.GT.NEDGO TO 2505
XK TO SEE IF THIS IS A FITTING SPECIES ALREADY
IF(ELFT(IEL,SIZ).E6.STAR)GO TO 2530
WRITE(UNIT6,2525)ELMENT(IEL)
FORMATMH ,A8,' IS NOT A FITTING SPECIES')
GO TO (2410,2510,2610,2710,3402) ,JRTN
MEL=MEL1(SIZ)
DO 2540 KEL=1,MEL
IF(ELHOLD(KEL,SIZ) .EB.IEDGO TO 2560
CONTINUE
GO TO 2523
MEL1 (SIZ)=MEL1(SIZ)-1
MEL=MEL1(SIZ)
ELFT(IEL,SIZ)=IBLANK
DO 2570 LEL=KEL,MEL
ELHOLD(LEL,SIZ)=ELHOLD(LEL+1,SIZ)
1 18
-------
TABLE A.1 CMB MAIN PROGRAM
(CONTINUED)
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
76 1
762
763
764
765
766
767
768
2570
C
C
C
C
C
2600
2605
2610
2615
2617
c CH:
2625
2630
2680
2690
C
c
c
c
c
2700
2705
2710
2715
2717
CONTINUE
GO TO (2410
2510,2610,2710,3402),JRTN
-AS.
JRTN=3
WRITE(UNIT6
GO TO 2610
WRITE(UNIT6
WRITE(UNIT6
FORMATUH ,
ADD A SOURCE TO THE FIT
31 10)SIZNAM(SIZ)
,2407)
,2615)
'INPUT
CODE OF ADDED SOURCE')
READ(UNIT5,2420,ERR=2605)JSOURC,SPACE
IF(SPACE.Eg.IBLANK)JSOURC=JSOURC/10
IF(JSOURC.E2.0)GO TO (500,2860,4120,4130,4132,4150,4250,4260,
*4270,4280),IRTN
IF( JSOURC. LT. 1 .OR. JSOURC. GT.NSOUROGO TO 2605
:CK TO SEE IF THIS IS ALREADY A FITTING SOURCE
IF(SOUFT(JSOURC,SIZ).EC.IBLANK)GO TO 2630
WRITE(UNIT6,2625)SOUNAM(JSOURC)
FORMATUH ,A8,' IS ALREADY A FITTING SOURCE')
GO TO (2410,2510,2610,2710,3402),JRTN
IF(MSORC1(SIZ).E6.MSOMAX)GO TO 2680
MSORC1(SIZ)=MSORC1(SIZ)+1
MSOURC=MSORC1(SIZ)
SOUFT(JSOURC,SIZ)=STAR
SOHOLD(MSOURC,SIZ)=JSOURC
GO TO (2410,2510,2610,2710,3402),JRTN
WRITE(UNIT6,2690)MSOMAX,SOUNAM(JSOURC)
FORMATUH ,'MAXIMUM OF ',12,' FITTING SOURCES EXCEEDED. ',
:1X,A8,' NOT ADDED TO FIT.')
GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
-DS. DELETE A SOURCE FROM THE FIT
JRTN=4
WRITE(UNIT6,3110)SIZNAM(SIZ)
GO TO 2710
WRITECUNIT6,2407)
WRITE(UNIT6,2715)
FORMATMH , 'INPUT CODE OF DELETED SOURCE1)
READ(UNIT5,2420,ERR=2705)JSOURC,SPACE
IF(SPACE.Eg.IBLANK)JSOURC=JSOURC/10
IF(JSOURC.E6.0)GO TO (500,2860,4120,4130,4132,4150
*4270,4280),IRTN
4250,4260
1 19
-------
TABLE A.I CMB MAIN PROGRAM (CONTINUED)
769 IFCJSOURC.LT.1.OR.JSOURC.GT.NSOURC)GO TO 2705
770 C IS IT ALREADY A FITTING SOURC
771 IFtSOUFTCJSOURC,SIZ).E2.STAR)GO TO 2730
772 2723 WRITE(UNIT6,2725)SOUNAM(JSOURC)
773 2725 FORMATMH ,A8,' IS NOT A FITTING SOURCE')
774 GO TO (2410,2510,2610,2710,3402),JRTN
775 2730 MSOURC=MSORC1(SIZ)
776 DO 2740 KSOURC=1,MSOURC
777 IF(SOHOLD(KSOURC,SIZ).E2.JSOUROGO TO 2760
778 2740 CONTINUE
779 GO TO 2723
780 2760 MSORC1(SIZ)=MSORC1(SIZ)-1
781 MSOURC=MSORC1(SIZ)
782 SOUFT(JSOURC,SIZ)=IBLANK
783 DO 2770 LSOURC=KSOURC,MSOURC
784 SOHOLD(LSOURC,SIZ)=SOHOLD(LSOURC+1,SIZ)
785 2770 CONTINUE
786 GO TO (2410,2510,2610,2710,3402), JRTN
787 C
788 C
789 C
790 C
791 C -CMB. PERFORM CHEMICAL ELEMENT BALANCE
792 2800 NOSKIP=0
793 2801 IF(NOSKIP.E2.1)GO TO 28032
794 WRITE(UNIT6,28000)
795 28000 FORMAT(' DO YOU WANT RIDGE SOLUTIONS?1)
796 READ(UNIT5,28001)ANS
797 28001 FORMAT(A1)
798 IF(ANS.E2.2UOTS)GO TO 28032
799 RIDGE=0
800 CWT= 1 . 0
801 IF(ANS.NE.YESY) GO TO 28020
802 RIDGE=1
803 WRITE(UNIT6,28010)
804 28010 FORMAT(? CMB WILL SELECT THE BEST RIDGE SOLUTION1/
805 * ' DO YOU WANT TO SEE A SUMMARY OF THE OTHERS?')
806 READ(UNIT5,28001)ANS
807 IF(ANS.E2.YESY)RIDGE=2
808 WRITE(UNIT6,28015)
809 28015 FORMAT(' IN THE SELECTION OF THE BEST RIDGE SOLUTION,
810 * /' WHAT WEIGHT DO YOU WANT? (DEFAULT=1.0)')
811 READ(UNIT5,28016)CWTO
812 28016 FORMAT(F10.0)
813 IF(CWTO.NE.0)CWT=CWTO
814 WRITE(UNIT6,28017)CWT
815 28017 FORMAT(' WEIGHT IS',F20.9)
316 28020 WRITE(UNIT6,28021 )
120
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
817 28021 FORMAT(' DO YOU WANT TO USE EFFECTIVE VARIANCES?')
818 EFVAR=0
819 READ(UNIT5,28001)ANS
820 IF(ANS.E2.YESY)EFVAR=1
821 2820 CONTINUE
822 WRITE(UNIT6,28031)
823 28031 FORMATC' DO YOU WANT TO INCLUDE THE INTERCEPT IN THE'
824 * ' MODEL?')
825 INTCEP=0
826 READCUNIT5,28033)ANS,KBUG
827 28033 FORMAT(A 1,2X,11)
828 IF(ANS.NE.YESY) GO TO 28032
829 INTCEP=1
830 IF(MSOURC.LE.MSOMIX) GO TO 28032
831 WRITE(UNIT6,28034) MSOMAX
832 28034 FORMATC/' NUMBER OF SOURCES PLUS INTERCEPT EXCEEDS ',12)
833 GO TO 500
834 28032 CONTINUE
835 C ENCODE TRANSFER ARRAYS
836 MEL=MEL1(SIZ)
837 MSOURC=MSORC1(SIZ)
838 1=0
839 DO 2830 KEL=1,MEL
840 IEL=ELHOLD(KEL,SIZ)
841 C(KEL)=UCNC(IEL,SIZ)
842 V(KEL)=UUCNC(IEL,SIZ)
843 IF(V(KEL).GT. 0.0)GO TO 2824
844 IF(I.E2.0)WRITE(UNIT6,2822)
845 2822 FORMAT(' **ERROR** SOME RECEPTOR CONCENTRATION STANDARD ERRORS'
846 * ' ARE LESS THAN'/' OR E2UAL TO ZERO. WEIGHTED REGRESSION '
847 * 'CANNOT BE DONE IN THIS CASE.')
848 WRITE(UNIT 6,2823)IEL,ELMENT(IEL),UCNC(IEL,SIZ),UUCNC(IEL,SIZ)
849 2823 FORMAT(' ',12,1X,A8,3X,'CONC=',F10.3,5X,'STD ERR=',F10.3)
850 1=1
851 2824 CONTINUE
852 DO 2825 KSOURC=1,MSOURC
853 JSOURC=SOHOLD(KSOURC,SIZ)
854 F(KEL,KSOURC)=A(IEL,JSOURC.SIZ)
855 SF(KEL,KSOURC)=UA(IEL,JSOURC.SIZ)
856 2825 CONTINUE
857 2830 CONTINUE
858 DO 2832 JSOURC=1,35
859 VIFELKCJSOURC)=0.0
860 SELK(JSOURC)=0.0
361 UELK(vTSOURC)=0. 0
862 2832 CONTINUE
863 IF(I.GT.O)GO TO 500
&G4 TOTALA=UCMSS
121
-------
TABLE A.1 CMS MAIN PROGRAM
(CONTINUED)
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
90 1
902
903
904
905
906
907
908
909
9 10
91 1
9 12
IFCSIZ.E8.1)TOTALA=UFMSS
CALL CMBR(F,C,MEL,MSOURC,MELMAX,MSOMAX,SF.V,VK,TOTALA,SSET,
* KINDEX,SAS,SAA,SR2,R2LS,VIF,
* RIDGE,EFVAR,INTCEP,IERR,KBUG,CWT,
* W,XP,YP,XMEAN,XPX,XPXK,XPY,SX,S,TM,AP,SOLD,IS,IA,ITM
* )
IFLAG(SIZ)=IERR
C INITIALIZE ALL SOURCES TO ZERO
2840 DO 2845 JSOURC=1,NSOURC
USORCCJSOURC,SIZ)=0.0
UUSORCCJSOURC,SIZ)=0.0
2845 CONTINUE
C REPLACE SOURCES WITH -CALCULATED VALUES
2850 CONTINUE
SETLS=SSETC1)
SETCSIZ)=SSETC2)
DO 2855 KSOURC=1,MSOURC
JSOURC=SOHOLDCKSOURC,SIZ)
USORCCJSOURC,SIZ)=SASCKSOURC,KINDEX)
UUSORCCJSOURC,SIZ)=SAACKSOURC,KINDEX)
VIFELKCJSOURC)=VIFCKSOURC>
SELKCJSOURC)=SASCKSOURC,1)
UELKCJSOURC)=SAACKSOURC,1)
2855 CONTINUE
IFCINTCEP.E2.0)GO TO 2856
I=MSOURC+1
SAINTCSIZ)=SASCI,KINDEX)
USAINTCSIZ)=SAACI,KINDEX)
SAINTO=SASCI,1)
USAINO=SAACI,1)
GO TO 2857
2856 SAINTCSIZ)=0.0
SAINTO=0.0
USAINTCSIZ)=0.0
USAINO^O.O
C GO TO CALCON
2857 IRTN=2
GO TO 3800
2860 IRTN=1
C PRINT ON SCREEN
C NO SOLUTION DISPLAYED FOR SINGULAR MATRIX
IFCIERR.ES.1)GO TO 500
GO TO 3200
C
C
C
C
COMMAND -PSOLN. PRINT ALL RIDGE SOLUTIONS ON SCREEN
122
-------
TABLE A.I CUB MAIN PROGRAM (CONTINUED)
913 C OR HARD COPY
914 C
915 42000 CONTINUE
916 IF(RIDGE.E2.0)GO TO 42060
917 J=31*(2*MSOURC+1)
918 I=MSOURC+1
919 WRITE(UNIT6,42001) MSOURC.J
920 42001 FORMAT(' THERE ARE 31 SOLUTIONS WITH AN INTERCEPT'/
921 * ' AND ',12,' SOURCE COEFICIENTS AND STANDARD ERRORS'/
922 * ' EACH FOR ',14,' NUMBERS TOTAL.'/
923 * ' THEY CAN PRINT TO YOUR TERMINAL OR TO HARD COPY.'/
924 * ' ARE YOU SURE YOU WANT TO LOOK AT THESE NUMBERS?')
925 READ(UNIT5,28001)ANS
926 IF (ANS.NE.YESY) GO TO 500
927 WRITE(UNIT6,42002)
928 42002 FORMATC' DO YOU WANT THE SOLUTIONS PRINTED ON THE HARD COPY
929 * / ' (INSTEAD OF AT YOUR TERMINAL)?')
930 READ(UNIT5,28001)ANS
931 UNITX=UNIT6
932 MANY=2
933 IF(ANS.NE.YESY)GO TO 42004
934 UNITX=UNIT11
935 MANY=6
936 GO TO 42005
937 42004 WRITE(UNIT6,42010)
938 42010 FORMAT(' CAN YOUR TERMINAL DISPLAY A 132 CHARACTER LINE?')
939 READ(UNIT5,28001)ANS
940 IF(ANS.E2.YESY)MANY=6
941 42005 MORE=MANY+1
942 IFdNTCEP. E6- 0 ) WRITE ( UNITS , 42011)
943 42011 FORMAT(/' ZERO INTERCEPT IN MODEL'/)
944 DO 42050 KEL=1,31,MORE
945 IEL=KEL+MANY
946 IFCIEL.GT.31)IEL=31
947 WRITE(UNITX,42006)(VK(J),J=KEL,IEL)
948 IFdNTCEP. GT. 0)
949 *WRITE(UNITX,42007)(SAS(I,J),SAA(I,J),J=KEL,IEL)
950 42006 FORMAT(//' RIDGE K' ,5X,7(F7.3 , 1 OX ) )
951 42007 FORMAT(' INTERCEPT',3X, 7(F7.3,'+-',F6.3,2X))
952 DO 42040 KSOURC=1,MSOURC
953 JSOURC=SOHOLD(KSOURC,SIZ)
954 WRITE(UNITX,42008) SOUNAM(JSOURC),(SAS(KSOURC,J),
955 * SAA(KSOURC,J),J=KEL,IEL)
956 42008 FORMAT(' ' ,A8,4X,7(F7 . 3, '+-',F6.3,2X))
957 42040 CONTINUE
958 42050 CONTINUE
959 IF (UNITX.E2.UNIT11 ) WRITE(UNIT6,4205 1 )
960 42051 FORMAT(' SOLUTIONS WRITTEN TO HARD COPY')
123
-------
TABLE A.I CMS MAIN PROGRAM
(CONTINUED)
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
99 1
992
993
994
995
996
997
998
999
1000
1C01
1002
1003
1004
1005
1006
1007
1 008
42060
42061
C
C
COMMA^
C
C
C
43000
C
43002
43005
43010
43020
43060
43039
43040
43310
C DIS
43200
43210
GO TO 500
WRITE(UNIT6,42061)
FORMATC' THE RIDGE
GO TO 500
ID -PMATRIX
CONTINUE
OPTION WAS NOT USED - ONLY 1 SOLUTION')
PRINT SOURCE SIGNATURE MATRIX, RECEPTOR
RECEPTOR CONCENTRATIONS
CODES FOR SOURCES AND SPECIES
COPY'/
YOUR TERMINAL
INSTEAD?')
OR CODES?')
LOOP=0
UNITX=UNIT11
WRITE(UNIT6,43002)
FORMAT(' MATRICES CAN GO TO HARD
* ' DO YOU WANT THEM DISPLAYED AT
READCUNIT5,28001) ANS
IF(ANS.E2.YESY)UNITX=UNIT6
WRITECUNIT6,43010)
FORMAT(' WHAT DO YOU WANT TO SEE?'/
* ' SOURCE SIGNATURE, RECEPTOR CONCENTRATIONS,
IF(LOOP.Eft.1)WRITE(UNIT6,43020)
FORMATC' OR ARE YOU DONE?')
READ(UNIT5,28001,ERR=43005) ANS
IF (ANS.E2.2UOTD .OR. ANS.E2.YESY .OR. ANS.E2.IBLANK)
* GO TO 500
LOOP=1
IF(ANS.E2.2UOTR)GO TO 3700
IF(ANS.E2.2UOTS)GO TO 43039
IF(ANS.E2.2UOTC)GO TO 43800
WRITE(UNIT6,43060)
FORMAT(' INVALID RE2UEST')
GO TO 43005
CONTINUE
WRITE(UNIT6,43040)
FORMAT(' WHAT SIZE FRACTION?
READ(UNIT5,28001)CHK
ISIZ=1
IF(CHK.E2.2UOTC) ISIZ=2
FORMAT(' ',12,1X,A8,5X,F8
PLAY SOURCE SIGNATURE
CONTINUE
WRITECUNIT6,43210) NSOURC.NEL
FORMAK' DO YOU WANT TO LOOK AT
* ' IT IS ',12,' SOURCES BY ',12
READ(UNITS,28001)ANS
IF(ANS.NE.YESY) GO TO 43500
(FINE OR COARSE)')
4,2X,
-' ,2X,F8.4)
THE WHOLE
, ' SPECIES
MATRIX?'/
' )
124
-------
TABLE A.I CMB MAIN PROGRAM
(CONTINUED)
1009
0 10
01 1
012
013
014
1015
1016
1017
10 18
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
C DISI
43045
43213
43218
43220
43221
43230
43240
C DIS
43500
43502
43510
43515
C DIS
43520
43800
43810
43815
43820
43830
A 132 CHARACTER LINE?')
IEL)
LAY WHOLE SOURCE SIGNATURE MATRIX
IWIDE=1
IF(UNITX.NE.UNIT6)GO TO 43213
WRITECUNIT6,43045)
FORMAT(f CAN YOUR TERMINAL DISPLAY
READ(UNITS,28001)ANS
IF(ANS.NE.YESY)IWIDE=0
MANY=5
IFCIWIDE.EB.1)MANY=12
MORE=MANY+1
DO 43240 KEL=1,NSOURC,MORE
IEL=KEL+MANY
IF(IEL.GT.NSOURC)IEL=NSOURC
WRITECUNITX,43218)(SOUNAM(I),I=KEL
FORMAT(' ',14X,13A9)
DO 43230 J=1,NEL
WRITE(UNITX,43220)J, ELMENTCJ),(ACJ,I,ISIZ),I=KEL,IEL)
WRITE(UNITX,43221) (UA(J,I,ISIZ),I=KEL,IEL)
FORMAT(/f ',12,1X,A8,3X,13F9.4)
FORMATC' ',14X,13F9.4)
CONTINUE
CONTINUE
GO TO 43005
'LAY INDIVIDUAL SOURCE SIGNATURES
CONTINUE
WRITE(UNIT6,43510)
FORMAT(' WHICH SOURCE DO YOU WANT?'/
* ' GIVE SOURCE t')
READ(UNITS,2420,ERR=43502)I,SPACE
IF(SPACE.Ee.IBLANK)I=I/10
IF(I.LE.O)GO TO 43005
WRITE(UNITX,43515) SOUNAM(I)
FORMAT(' SOURCE: ',A8)
DO 43520 J=1,NEL
WRITE(UNITX,43310) J,ELMENT(J),A(J,I,ISIZ),UA(J,I.
>LAY SOURCES AND SPECIES CODE LISTS
CONTINUE
GO TO 43500
CONTINUE
WRITECUNITX
ISIZ)
43810)
FORMATC/' SOURCE #',8X
DO 43820 I=1,NSOURC
WRITECUNITX,43815) I,SOUNAM(I)
FORMATC' ',3X.I2,10X
CONTINUE
WRITECUNITX,43830)
FORMATC/1 SPECIES f
DO 43840 1=1,NEL
NAME',12X,'CODE')
SCODECI)
A8, 10X,I2)
7X,'NAME',12X,'CODE')
125
-------
TABLE A.I CMB MAIN PROGRAM
(CONTINUED)
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1 092
1093
1094
1095
1096
1097
1098
1099
1 100
1101
1 102
1 103
1 104
43840
43850
*
*
C
c
C
c
3000
3005
3030
*
*
*
3035
*
*
3040
3045
3055
3060
3080
C
C
c
c
c
c
c
3100
31 10
WRITE(UNITX,43815)I,ELMENT(I>,PCODE(I)
CONTINUE
WRITECUNITX,43850)
FORMATC/' USE SOURCE * AND SPECIES t FOR ALL YOUR
'RESPONSES'/ ' USE CODES ONLY WHEN CODING '
'MATRICES FOR THE DATA BASE')
GO TO 43005
-PINFO. PRINT CURRENT STATUS ON SCREEN
WRITECUNIT6,3005)
FORMATC' ***CURRENT STATUS***1)
WRITE(UNIT6,3030)CMBID,SIZNAM(SIZ),UDUR,USTART,YBACKCIBACK)
FORMATC' CMB SITE: ' , 3A4 , 5X , ' YEAR -' ' , A2 , 3X , ' DATE : ',A4/
IX,A8,'SIZE FRACTION'/
*1X,'DURATION: ',12,' START HOUR: »,I2,/,
*1X,'HAS BACKGROUND BEEN SUBTRACTED: ',A3)
IFCIBACK.ES.2)GO TO 3040
WRITE(UNIT6,3035)BACKID,BDUR,BSTART
FORMATC1H ,'SUBTRACTED BACKGROUND CMB SITE: ',3A4,5X,
'YEAR: ' , A2.3X,'DATE'- ',A4/
*1X,'DURATION: ',12,* START HOUR: ',12)
URITE(UNIT6,3045)
FORMATC1H ,'FITTING SPECIES')
DO 3055 KEL=1,MEL
IEL=ELHOLD(KEL,SIZ)
WRITECUNIT6,425)IEL,ELMENTCIEL)
CONTINUE
WRITECUNIT6.3060)
FORMATC1H ,'FITTING SOURCES')
DO 3080 KSOURC=1,MSOURC
JSOURC=SOHOLDCKSOURC,SIZ)
WRITECUNIT6,425)JSOURC,SOUNAMCJSOURC)
CONTINUE
GO TO C500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
-SIZE.
CHANGES SIZE FRACTION BETWEEN FINE AKD COARSE
ISIZ=SIZ
IFCISIZ.EB- 1 )SIZ = 2
IFCISIZ.E2.2)SIZ=1
WRITECUNIT6,3110)SIZNAMCSIZ)
FORMATC' SIZE IS ',A8)
126
-------
TABLE A.1 CMS MAIN PROGRAM
(CONTINUED)
1 105
1 106
1 107
1 108
1 109
1110
1111
1112
1113
1 1 14
1115
1116
1 1 17
1 1 18
1119
1 120
1121
1 122
1 123
1 124
1 125
1 126
1 127
1 128
1 129
1 130
1131
1 132
1 133
1 134
1 135
1 136
1 137
1 138
1 139
1 140
1141
1 142
1 143
1 144
1 145
1 146
1 147
1 148
1 149
1 150
1151
1 152
IRTN=1
GO TO 500
C
C
C
C
C
C
C
-PDATA. PRINTS CURRENT CMB RESULTS TO SCREEN
3200 I=2-EFVAR
WRITE(UNIT6,3205)CMBID,SIZNAM(SIZ),UDUR,USTART,YBACK(IBACK),
* YBACK(I),SR2,VK(KINDEX),R2LS
3205 FORMAT('OCMB SITE: • , 3A4 , 5X , ' YEAR •' ' , A2 , 3X , ' DATE : ',A4,6X,
* 'FRACTION: ',A6/ ' SAMPLE DURATION' ',12,
* ' WITH START HOUR: ',12,2X,'BACKGROUND: ',A3,
* 2X,'EFF VAR: ',A3/
* 1X,9('-'),'RIDGE REGRESSION',8('-'),4X,5('-'),
* 'WEIGHTED LEAST SQUARES',5('-')/
* 3X,'R-S2UARE:',F6.4,3X,'RIDGE K=',F5.3,12X.
* 'R-S6UARE:',F6.4/
* ' SOURCE',10('-'),'UG/M3',9('-'),4X,7('-'),'UG/M3',
* 1 0('-'),'VIF',7('-'))
WRITE(UNIT6,3208)SAINT(SIZ),USAINT(SIZ),SAINTO,USAINO
3208 FORMAT(3X,'INTERCEPT',F10.3,'+-',F8.3,5X,F10.3,'+-',F8.3)
X=SAINT(SIZ)
XX=SAINTO
DO 3240 JSOURC=1.NSOURC
IF(UUSORC(JSOURC.SIZ).LE.0.)GO TO 3240
WRITE(UNIT6,3230)JSOURC,SOUNAM(JSOURC),
1USORC(JSOURC.SIZ),UUSORC(JSOURC,SIZ),SELK(JSOURC),UELK(JSOURC),
* VIFELK(JSOURC)
3230 FORMATdH , 12 , 1 X , A8 , F 1 0 . 3 , ' +- ' , F8 . 3 ,
* 5X.F10.3,'+-',F8.3,1X.F8.3)
X=X+USORC(JSOURC.SIZ)
XX=XX+SELK(JSOURC)
3240 CONTINUE
WRITE(UNIT6,3245)X,SET(SIZ),XX,SETLS
3245 FORMATMH , 'TOTAL: ' , 4X , F 1 0 . 3 , ' +- ' , F8 . 3 , 5X , F 1 0 . 3 , ' +- ' , F8 . 3 )
WRITE(UNIT6,3279)
WRITE(UNIT6,3280)UFMSS,UUFMSS,UCMSS,UUCMSS,UTMSS,UUTMSS
3279 FORMATMH .'MEASURED CONCENTRATION FINE/COARSE/TOTAL:')
3280 FORMATdH , 2 ( F 1 0 . 5 , ' +- ' , F8 . 3 , ' / ' ) , F 1 0 . 5 , ' + - ' , F8 . 3 )
3241 WRITE(UNIT6,3250)
3250 FORMAT(1X,'PRESS RETURN TO CONTINUE OR ENTER C FOR NEXT ',
* 'COMMAND')
3243 READ(UNITS,28001,ERR=3248)ANS
IF(ANS .E2. 2UOTOGO TO 3275
IF(ANS .E2. IBLANK)GO TO 3244
127
-------
TABLE A.I CMB MAIN PROGRAM (CONTINUED)
1153 C OPTION TO "STOP" IN AUTOFIT
1154 IF(ANS .HE. 2UOTS)GO TO 3248
1155 3246 IRTN=1
1156 KCNTRL=ICNTRL
1157 ICNTRL=4
1158 GO TO 500
1159 3248 WRITE(UNIT6,3249)
1160 3249 FORMAK 1X, 'INPUT ERROR.')
1161 GO TO 3241
1162 3244 WRITE(UNIT6,3247)
1163 3247 FORMATUH ,25('-'),'MEAS',12('-'),'CALC',10('-'),'RATIO«,10(•-•))
1164 CORRECT STD. ERR. FOR TOTAL AEROSOL
1165 UCLCN(1,SIZ)=SETCSIZ)
1166 DO 3270 IEL=1,NEL
1 167 IFCUCNC(IEL,SIZ).LE.0 . )RATIO = 0 .
1168 IF(UCNC(IEL,SIZ).LE.0.)URATIO=0.
1169 IF(UCNC(IEL,SIZ).LE.0)GO TO 3242
1170 RATIO=CLCN(IEL,SIZ)/UCNC(IEL,SIZ)
1171 URATIO=((UCLCN(IEL,SIZ)/UCNC(IEL,SIZ))**2+(UUCNC(IEL,SIZ)
1172 1*CLCN(IEL,SIZ)/UCNC(IEL,SIZ)**2)**2)**.5
1173 3242 IF(UCNC(IEL,SIZ). GE.UUCNC(IEL,SIZ))
1174 1WRITE(UNIT6,3260)IEL,ELMENT(IEL),ELFT(IEL,SIZ),MFLAG(IEL,SIZ),
1175 1UCNC(IEL,SIZ),
1176 1UUCNC(IEL,SIZ),CLCN(IEL,SIZ),UCLCNdEL,SIZ),RATIO,URATIO,
1177 1ELMENT(IEL)
1178 IF(UCNC(IEL,SIZ).LT.UUCNC(IEL,SIZ))
1179 1WRITE(UNIT6,3265)IEL,ELMENT(IEL),ELFT(IEL,SIZ),MFLAG(IEL,SIZ),
1 180 1UUCNC(IEL,SIZ),CLCN(IEL,SIZ),UCLCN(IEL,SIZ),
1181 1RATIO,URATIO,ELMENT(IEL)
1182 3260 FORMATMH , 12 , 1 X , A8 , 2X , A 1 , 2X , A 1 , 1 X , F7 . 3 , ' +- ' , F6 . 3 , 1 X ,
1183 11X,F7.3,'+-',F6.3,1X,F5.2,'+-',F5.2,2X,A3)
1184 3265 FORMATMH , 12 , 1X , A8 , 2X , A 1 , 2X , A 1 , 1 X , 7X , ' < ',F6.3,1X,
1185 11X,F7.3,'+-',F6.3,1X,F5.2,'+-',F5.2,2X,A3)
1186 3270 CONTINUE
1187 CHANCE TO HOLD FOR CRT
1188 IF(ICNTRL.NE.2 .AND. ICNTRL.NE.0)GO TO 3276
1189 IRTN=5
1190 IF(SIZ.ES.2 .AND. ICNTRL.Eg.0)GO TO 3276
1191 READ(UNIT5,28001)ANS
1192 IF(ANS.E2.eUOTS)GO TO 3246
1193 GO TO 3276
1194 3275 IF(ICNTRL.E2.2 .OR. ICNTRL.E2.0)IRTN=5
1195 3276 GO TO (500,2860 , 4120 , 4130,4132,4150,4250,4260,4270,4280),IRTN
1196 C
1197 C
1 198 C
1199 C
1200 C -WRITE. WRITES DATA TO HARDCOPY
128
-------
TABLE A.1 CMB MAIN PROGRAM
(CONTINUED)
1201
1202
I 203
1204
1205
1206
1207
1208
1209
12 10
1211
1212
1213
1214
1215
1216
1217
12 18
12 19
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
3400
C
3402
JRTN=5
3413
3415
3417
3430
3440
3450
3460
3468
UNIT=UNIT11
ISIZ=SIZ
IPF=IFLAG(ISIZ)
WRITE(UNIT,3415)CMBID,(FLAG(I,IPF +1),1=1,8),SIZNAIK ISIZ)
FORMAT(' ',/,1X,'RESULTS FOR CMB SITE: ',3A4,5X,
* 'YEAR: ' ,A2,3X, 'DATE-* ',A4,4X,8A4,
*1X,/,f ',A6,f PARTICULATE FRACTION')
FORMAT(1X,' ')
WRITE(UNIT,3430)UDUR,USTART
FORMAT(
*1X,'SAMPLING DURATION: ',12,' HRS. WITH START HOUR: ',12)
WRITE(UNIT,3440)YBACK(IBACK)
FORMATMH , 'BACKGROUND SITE SUBTRACTED' ',A3)
IF(IBACK.E2.2)GO TO 3460
WRITE BACKG INFO
WRITE(UNIT,3450)BACKID
FORMAT(' BACKGROUND CMB SITE:
* 'YEAR: ',A2,3X,'DATE: ',A4)
WRITE(UNIT,3430)BDUR,BSTART
IFdSIZ.NE. 3)GO TO 3466
WRITE(UNIT,3468)
FORMAT(' RESULTS DERIVED FROM
SAINT(3)=SAINT(1)+SAINT(2)
USAINT(3)=S2RT(USAINT( 1 ) **2+USAINT(2)**2)
WRITE(UNIT,3477)SAINT(3),USAINT(3)
GO TO 3478
IF(INTCEP.EQ.O)WRITE(UNIT,3469)
FORMAT(f INTERCEPT SET TO ZERO')
IF(RIDGE.Eg.0)WRITE(UNIT,3470)
RIDGE OPTION NOT USED')
GT.O)WRITE(UNIT,3471)
EFFECTIVE VARIANCE METHOD')
',3A4,5X,
FINE AND COARSE FITTINGS')
3466
3469
3470
3471
FORMAT('
IF(EFVAR.
FORMAT('
C WRITE HEADER
3474 WRITE(UNIT,3476)VK(KINDEX),SR2,R2LS,SAINT(ISIZ),USAINT(ISIZ)
* SAINTO.USAINO
3476 FORMAT(25X,'RIDGE REGRESSION',25X,'LEAST SgUARES'/
*
*
*
0',5X,
25X, 'K=' ,F5.3,5X, 'R-SCUARE: ' ,F6.4, 13X, 'K = 0
'R-SGUARE: ',F6.4/' ',120('-')/
1X, 'CODE' , 2X, 'SOURCE1 , 1 3X , 'UG/M3* , 15X, ' 5J' ,
* 22X,'UG/M3',13X,'VIF'/' ',120('-')
* /5X,'INTERCEPT1, 3X,F8.3,'+-',F8.3,26X,F8.3,'+-',F8.3)
3477 FORMAT(/' ', 1 20('-')/1X,'CODE',2X,'SOURCE', 13X,'UG/M3',
* 15X, '?.'/" ', 120('-' )/5X, 'INTERCEPT',3X,F8.3, '+-' ,F8.3)
C WRITE SOURCES
3478 TX=SAINT(ISIZ)
XX=SAINTO
129
-------
TABLE A.I CMB MAIN PROGRAM
(CONTINUED)
12'19
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
126(4
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1 290
1291
1292
1293
1294
1295
1296
. AND.UUSORC(JSOURC, 2)
I5CTR=0
DO 3490 JSOURC-1.NSOURC
IF(UUSORC(JSOURC,1 ) .LE.0
1 .LE.0. )GO TO 3490
IF(ISIZ.E2.1.AND.UUSORC(JSOURC,1).LE.0
1)X=UFMSS
1)UX=UUFMSS
2)X=UCMSS
2)UX=UUCMSS
3)X=UTMSS
,3)UX=UUTMSS
.Eg.
,E2.
,E2.
,E2.
. E2
.E2
3480
3485
)GO TO 3490
IFdSIZ.
IFCISIZ.
IFdSIZ,
IFdSIZ.
IFdSIZ
IFdSIZ
GET TOTAL SOURCE
IFCISIZ.NE.3)GO TO 3480
USORC(JSOURC,3)=USORC(JSOURC,1)+USORC(JSOURC,2)
UUSORC(JSOURC,3)=((UUSORC(JSOURC,1))**2+UUSORC(JSOURC,2)**2)**.5
CALL PERC(PCNT,UPCNT,USORC(JSOURC,ISIZ),UUSORC(JSOURC,ISIZ),X,UX)
IFCUX.LE.0. )GO TO 3490
IFdSIZ.LT.3)WRITE(UNIT,3485)JSOURC,SOUNAM(JSOURC),
1USORC(JSOURC,ISIZ),UUSORC(JSOURC,ISIZ),PCNT,UPCNT,
* SELK(JSOURC),UELK(JSOURC),VIFELK(JSOURC)
IF(ISIZ.E2.3) WRITE(UNIT,3485) JSOURC,SOUNAIK JSOURC),
* USORC(JSOURC,ISIZ),UUSORC(JSOURC,ISIZ),PCNT,UPCNT
FORMAT(1X,I2,3X,A8,1X,F10.3,'+-',F8.3,2X,F7.3,'+-',F6.3,
* 7X,F10.3,'+-',F8.3,2X,F8.3)
TX=TX+USORC(JSOURC,ISIZ)
XX=XX+SELK(JSOURC)
INCREMENT SOURCE COUNTER
ISCTR=ISCTR+1
TOTAL FRACTION, HOLD SOURCE CODES FOR SUMMARY
IF (ISIZ.E2.3)ISHOLD(ISCTR)=JSOURC
CONTINUE
IF(ISIZ.E2.3)SET(3)=S2RT(SET(1)**2+SET(2)**2)
WRITE(UNIT,3499)
CALL PERC(PCNT,UPCNT,TX,SET(ISIZ),X,UX)
IFdSIZ.LT.3)WRITE(UNIT,3491)TX,SET(ISIZ),PCNT,UPCNT,XX,SETLS
IF(ISIZ.E2.3)WRITE(UNIT,3491)TX,SET(ISIZ),PCNT,UPCNT
FORMAT(6X,'TOTAL:',3X,F10.3,'+-',F8.3,2X,F7.3,'+-',F6.3,7X,F10.3,
* '+-' ,F8.3)
WRITE(UNIT,3417)
C WRITE ELEMENTS
C WRITE HEADER
WRITE(UNIT,3495)SIZNAM(ISIZ)
FORMAT(IX,'SPECIES FIT',2X,'MISS',9X,A6
*'SUSPENDED PARTICULATE')
WRITE(UNIT,3497)
FORMAT(' CODE FLG' ,2X, 'FLG' , 3X , 'MEAS .
*1X, 'PERCENT' ,7X, 'CALC. UG/M3' ,7X, 'RATIO1 )
WRITECUNIT,3499)
C IF
3490
3491
3495
3497
IX,
UG/M3',8X,
130
-------
TABLE A.1 CMB MAIN PROGRAM (CONTINUED)
1297 3499 FORMAT(1X,120('-'))
1298 3500 FORMAT(2X,I2,2X,A6,2X,A1,2X,A1,2X,2(F7.3,'+-',F6.3,1X),1X,
1299 1 1X.F7.3, '+-' ,F6.3, 1X,F6.3, '+-',F6.3,2X,A8)
1300 3505 FORMAT(2X,I2,2X,A6,2X,A1,2X,A1,2X,2(7X,'< ",F6.3,1X),
1301 12X,F7.3,'+-',F6.3,1X,F6.3,'+-',F6.3,2X,A8)
1302 CORRECT STD. ERR. FOR TOTAL
1303 UCLCN(1,ISIZ)=SET(ISIZ)
1304 DO 3530 IEL=1,NEL
1305 IF(ISIZ.E2.3)CLCN(IEL,ISIZ)=CLCN(IEL,1)+CLCN(IEL,2)
1306 IF(ISIZ.E2.3)UCLCN(IEL,ISIZ)=
1307 *SfiRT((UCLCNdEL,1)) **2+(UCLCN(IEL,2))**2)
1308 3525 IF(UCNC(IEL,ISIZ).E2.0.)RATIO=0.
1309 IF(UCNC(IEL,ISIZ).LE.0.)URATIO=0.
1310 IF(UCNC(IEL,ISIZ).LE.0.)GO TO 3526
1311 RATIO=CLCN(IEL,ISIZ)/UCNC(IEL,ISIZ)
1312 URATIO=S2RT((UCLCN(IEL,ISIZ)/UCNC(IEL,ISIZ))**2 +
1313 *(UCLCN(IEL,ISIZ)*CLCN(IEL,ISIZ)/UCNC(IEL,ISIZ)**2)**2)
1314 3526 CALL PERC(PCNT,UPCNT,UCNC(IEL,ISIZ),UUCNC(IEL,ISIZ),X,UX)
1315 IF(UCNC(IEL,ISIZ).GE.UUCNC(IEL,ISIZ))WRITE(UNIT,3500)IEL,
1316 *ELMENT(IEL),ELFT(IEL,SIZ),MFLAG(IEL,ISIZ),UCNC(IEL,ISIZ),UUCNC
1317 *(IEL,ISIZ),PCNT,UPCNT,CLCN(IEL,ISIZ),UCLCN(IEL,ISIZ),RATIO,URATIO,
1318 *ELMENT(IEL)
1319 IF(UCNC(IEL,ISIZ).LT.UUCNC(IEL,ISIZ))WRITE(UNIT,3505)IEL,
1320 *ELMENT(IEL),ELFT(IEL,SIZ),MFLAG(IEL,ISIZ),UCNC(IEL,ISIZ),UPCNT,
1321 *CLCN(IEL,ISIZ),UCLCN(IEL,ISIZ),RATIO,URATIO,ELMENT(IEL)
1322 3530 CONTINUE
1323 URITE(UNIT,3499)
1324 C WRITE MASSES
1325 WRITE(UNIT,3540)UFMSS,UUFMSS,UCMSS,UUCMSS,UTMSS , UUTMSS
1326 3540 FORMAT(' MEASURED AMBIENT MASS (UG/M3): FINE' ',
1327 *1X,F5. 1, '+-' ,F4.1, ' COARSE: ',F5.1,'+-',F4. 1,' TOTAL: ',
1328 *1X,F5. 1, '+-' ,F4. 1)
1329 WRITE(UNIT,3543)
1330 3543 FORMAT(/V//1 1')
1331 3545 WRITECUNIT6,3546)
1332 3546 FORMAK1 WRITTEN')
1333 IFdSIZ.NE. 1 )GO TO 3549
1334 DO 3547 IEL=1,5
1335 3547 OLDID(IEL)=CMBID(IEL)
1336 3548 GO TO (500,2860,4120,4130,4132,4150,4250,4260,
1337 *4270,4280),IRTN
1338 3549 IFCISIZ.Efi.3)GO TO 3600
1339 3550 DO 3552 IEL=1,5
1340 IF(OLDIDdEL) . NE . CMBID ( IEL ) ) GO TO 3548
1341 3552 CONTINUE
1342 ISIZ=3
1343 IPF = 0
1344 IF(IFLAG(1).Efi.2.OR.IFLAG(2).EQ.2)IPF=2
131
-------
TABLE A.I CUB MAIN PROGRAM (CONTINUED)
1345 IFdFLAGC 1 ) . EC. 1 .OR.IFLAG(2) . E2. 1 )IPF=1
1346 GO TO 3413
1347 C WRITE SUMMARY FILE
1343 3600 CONTINUE
1349 UNIT=UNIT9
1350 C WRITE HEADERS
1351 WRITE(UNIT,3605)
1352 3605 FORMAT(f '//39X,'*** CMS SOURCE CONTRIBUTION SUMMARY ***'/' ')
1353 3613 WRITE(UNIT,3615)CMBID
1354 3615 FORMATC1 f,/>36X,'RESULTS FOR CMB SITE: ',3A4,5X,
1355 * 'YEAR: ' , A2 , 3X , ' DATE '• ',A4)
1356 3617 FORMATUX,' ')
1357 WRITE(UNIT,3630)UDUR,USTART
1358 3630 FORMAT(
1359 *1X,SOX,'SAMPLING DURATION: ',12,' MRS. WITH START HOUR: ',12)
1360 WRITE(UNIT,3640)YBACK(IBACK)
1361 3640 FORMATC30X,'BACKGROUND SITE SUBTRACTED: ',A3)
1362 IF(IBACK.E2.2)GO TO 3652
1363 C WRITE BACKG INFO
1364 WRITECUNIT,3650)BACKID
1365 3650 FORMAT(30X,'BACKGROUND CMB SITE: ',3A4,5X,'YEAR: ',A2,3X,
1366 * 'DATE: ',A4)
1367 WRITECUNIT,3630)BDUR,BSTART
1368 3652 CONTINUE
1369 IS1=IFLAG<1)+1
1370 IS2=IFLAG(2)+1
1371 WRITE(UNIT,3654)1STAR(IS 1),ISTAR(IS2)
1372 3654 FORMAT(/,1X,' SOURCE ',18X,'FINE',A4,26X,'COARSE',A4,25X,'TOTAL'/
1373 *1X,117('-')/3(16X,'UG/M3',15X,'%'))
1374 C CALCULATE PERCENTS AND WRITE CONTRIBUTIONS
1375 FX=SAINT(1)
1376 CX=SAINT(2)
1377 TX=SAINT(3)
1378 WRITE(UNIT,3657)SAINT(1),USAINT(1),SAINT(2),USAINT(2),SAINT(3),
1379 * USAINTC3)
1380 3657 FORMAT(' INTERCEPT',3(1X,F7.3,' +-',F7.3, 1 8X ) )
1381 DO 3660 I=1,ISCTR
1382 J=ISHOLD(I)
1383 CALL PERC(FPCNT,UFPCNT,USORC(J,1),UUSORC(J,1),UFMSS,UUFMSS)
1384 CALL PERC(CPCNT,UCPCNT,USORC(J,2),UUSORC(J,2),UCMSS,UUCMSS)
1385 CALL PERC(TPCNT,UTPCNT,USORC(J,3),UUSORC(J,3),UTMSS,UUTMSS)
1386 WRITE(UNIT,3658)SOUNAM(J),USORC(J,1),UUSORC(J,1),FPCNT,UFPCNT,
1387 *USORC(J,2),UUSORC(J,2),CPCNT,UCPCNT,USORC(J,3>,UUSORC(J,3),TPCNT,
1388 *UTPCNT
1389 3658 FORMAT(1X,A8,1X,6(1X,F7.3,' +-',F7.3>)
1390 C WRITE FILE FOR PCOMP COMMAND TO READ
1391 WRITE(UNIT 10,3659)CMBID,J,SOUNAM(J),USORC(J, 1),UUSORC(J, 1 ) ,
1392 * USORC(J,2),UUSORC(J,2),USORC(J,3),UUSORC(J,3)
132
-------
TABLE A.I CUB MAIN PROGRAM
(CONTINUED)
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1 440
365'
C INi
3660
c CA:
3664
3671
3672
C WR
C CA
367
C WR
3677
3678
3680
3995
C
C
C
C
C
3659 FORMAT(3A4,A2,A4,1X,I2,1X,A8,2X,6F8.3)
INCREMENT SUMS
FX=FX+USORC(J, 1 )
CX=CX+USORC(J,2)
TX=TX+USORC(J,3)
CONTINUE
CALCULATE AND WRITE SUMS
CALL PERC(FPCNT,UFPCNT,FX,SET(1),UFMSS,UUFMSS)
CALL PERC(CPCNT,UCPCNT,CX,SET(2),UCMSS,UUCMSS)
CALL PERC(TPCNT,UTPCNT,TX,SET(3),UTMSS,UUTMSS)
WRITE(UNIT,3664)FX,SET(1),FPCNT,UFPCNT,CX,SET(2),CPCNT,UCPCNT,
*TX,SET(3),TPCNT,UPCNT,UFMSS,UUFMSS,UCMSS,UUCMSS,UTMSS,UUTMSS
FORMAT(1X,117('-')/,1X,'CALC.MASS1,6(1X,F7.3,' +-',F7.3)/
*1X,'MEAS.MASS',1X,3(F7.3,' +-',F7.3,19X)///)
IF(IFLAG(1) .Efi. 1.OR.IFLAG(2) .E2. 1)WRITE(UNIT,3671)(FLAG(1,2),1= 1 ,8
*>
IF(IFLAG(1).E2.2.OR.IFLAG(2).E2.2)WRITE(UNIT,3672)(FLAG(I,3),1=1,8
*>
FORMAT(X/,' *NOTE= f,8A4)
FORMAT(X/,' **NOTE-' ',8A4)
WRITE MAGSTO FILE
CALCULATE INDIVIDUAL ELEMENTAL CONTRIBUTIONS FOR EACH SOURCE
3676 DO 3680 I=1,ISCTR
J=ISHOLD(I)
WRITE MASS CONTRIBUTION
DO 3680 IEL=1,NEL
DO 3677 ISIZ=1,2
IF(A(IEL,J,ISIZ).LE.O.OE-09)SEL(ISIZ)=0.0
IF(A(IEL,J,ISIZ).LE.0.OE-09)USEL(ISIZ)=0.0
IF(A(IEL,J,ISIZ).LE.0.OE-09)GO TO 3677
SEL(ISIZ)=A(IEL,J,ISIZ)*USORC(J,ISIZ)
USEL(ISIZ)=S2RT(SEL(ISIZ)**2*((UA(IEL,J,ISIZ)/A(IEL,J,ISIZ))
***2+(UUSORC(J,ISIZ)/USORC(J,ISIZ))**2)
* + (UA(IEL,J,ISIZ)*UUSORC(0,ISIZ))**2)
CONTINUE
IF(SEL(1).E2.0.0.AND.SEL(2).E2.0.0)GO TO 3680
WRITE(UNIT12,3678)CMBID,UDUR,USTART,SCODE(J),PCODE(IEL)
*,SEL(1),USEL(1),SELC2),USEL(2)
FORMAT('40',1X,3A4,1X,A2,A4,1X,I2,1X,I2,1X,I2,1X,
*I2,2X,F9.4,2X,F9.4,2X,F9.4,2X,F9.4)
CONTINUE
GO TO(500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
133
-------
TABLE A.I CUB MAIN PROGRAM
(CONTINUED)
144 1
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
147-1
1472
1 473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1 487
1 488
C
C
C
C
C
3700
3710
3720
3730
C
C
C
C
C
C
3800
3840
3850
C
C
C
C
C
4000
C
C
C
C
C
409
4100
PART OF COMMAND PMATRIX
PRINTS CURRENT FINE AND COARSE MEASURED CONCENTRATIONS
WRITE(UNITX,3710)CMBID
FORMATUH ,'CURRENT MEASURED UG/M3 CONCENTRATIONS FOR'/' SITE:',
* 3A4,5X,'YEAR: ',A2,2X,'DATE= »,A4,
*/1X,71('-'),/,1H ,'SPECIES', 14X, 'FINE' , 13X, 'COARSE' , 13X,'TOTAL' ,
*/1X,71('-')//)
DO 3730 1=1,NEL
WRITE (UNITX, 3720) I, ELMENT CD , UCNCd, 1 ) , UUCNCd, 1 ) ,UCNC(I,2) ,
*UUCNC(I,2),UCNC(I,3),UUCNC(I,3)
FORMAT(1H ,I2,1X,A8,1X,F8.3,'+-',F8.3,2MX,F8.3,'+-',F8.3))
CONTINUE
GO TO 43005
-CALCON. CALCULATE CONCENTRATIONS.
DO 3850 IEL=1,N£L
CLCN(IEL,SIZ)=SAINT(SIZ)
UCLCN(IEL,SIZ)=0.0
KSOURC=NSOURC
DO 3840 JSOURC=1,KSOURC
IF(UUSORCCJSOURC,SIZ).LE.0.)GO TO 3840
CLCN(IEL,SIZ)=CLCN(IEL,SIZ)+A(IEL,JSOURC,SIZ)*USORC(JSOURC,SIZ)
UCLCN(IEL,SIZ)=(UCLCN(IEL,SIZ)**2+(UA(IEL,JSOURC,SIZ)*USORC
1 (JSOURC,SIZ))**2)**.5
CONTINUE
CONTINUE
GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
-RESUM. GO BACK INTO AUTOFIT
ICNTRL=KCNTRL
GO TO 4132
-AUTOFIT.
9 NOSKIP=0
CONTINUE
STEPS THROUGH COMMAND SE2UENCE FOR A GIVEN SAMPLE
134
-------
TABLE A.1 CMS MAIN PROGRAM (CONTINUED)
1489 C IF END OF AUTOFIT DATA, WRITE TO SCREEN AND GO TO COMMAND
1490 IFCICNTRL.NE.0)GO TO 4106
1491 WRITE(UNIT6,4105)
1492 4105 FORMATMH ,'AUTOFIT SEQUENCING FINISHED')
1493 ICNTRL=4
1494 IRTN=1
1495 GO TO 500
1496 4106 CONTINUE
1497 C IF FIRST SAMPLE GO TO SELECT
1498 IRTN=3
1499 IFCICNTRL.NE.2)GO TO 4210
1500 C RETRIEVE NEXT CMBID FROM PREVIOUS DATA SEARCH
1501 DO 41 10 1=1,5
1502 4110 CMBID(I)=SAVE(I)
1503 C CALL UP DATA
1504 4120 ICNTRL=1
1505 URITECUNIT6,4122)CMBID
1506 4122 FORMATMH ,'DATA SEARCH BEGUN FOR',/6X,'SITE: ',3A4,5X,
1507 * 'YEAR; •,A2.2X,'DATE: «,A4/)
1508 CALL FETCHC2UOTA,CMBID,UDUR,USTART,UCNC,
1509 *UUCNC,UFMSS,UUFMSS,UCMSS,UUCMSS,UTMSS,UUTMSS,MFLAG,
1510 *SAVE,ICNTRL,TYPE,UNIT6,UNIT12,UNIT13)
1511 IFdCNTRL.ES. 9)IRTN=1
1512 IF(ICNTRL.E2.9)GO TO 500
1513 4123 SIZ=1
1514 4124 CONTINUE
1515 C PERFORM CEB
1516 4130 IRTN=5
1517 GO TO 2801
1518 4132 NOSKIP=1
1519 C RETURN FROM RESUM- WRITE DATA, GO TO NEXT SIZE
1520 4135 IRTN=6
1521 GO TO 3400
1522 4150 CONTINUE
1523 CHECK IF DONE BOTH SIZES £ RECYCLE AUTOFIT
1524 IFCSIZ.E2.2)GO TO 4100
1525 SIZ=2
1526 GO TO 4124
1527 C
1528 C
1529 C
1530 C -SELECT. ALLOWS DATA SET SELECTION
1531 C
1532 4200 CONTINUE
1533 IRTN=7
1534 4210 WRITE(UNIT6,4215)
1535 4215 FORMATC' ENTER DESIRED CMB SITE CODE: XXXXXXXXXXXX')
1536 READ(UNIT5,4218,ERR=4210)SITE
135
-------
TABLE A.1 CUB MAIN PROGRAM
(CONTINUED)
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
4216
4217
4219
4220
4218
4250
C
4254
4255
C DO
4260
4270
4280
C
C
C
C
C
4400
4401
4410
4420
4430
4440
WRITE(UNIT6,4216)
READCUNIT5,4217) YEAR
WRITECUNIT6,4219)
READ(UNIT5,4220) DATE
FORMAT(' ENTER YEAR: YY')
FORMATCA2)
FORMATC' ENTER DATE' MMDD')
FORMAT(A4)
FORMAT(3A4)
GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
IRTN=8
WRITE(UNIT6,4255)
FORMATC' INPUT DESIRED SIZE FRACTION:(FINE OR COARSE)')
READ(UNIT5,28001,ERR=4254)ANS
SIZ=1
IF(ANS.EB.eUOTC) SIZ=2
ICNTRL=4
WRITE(UNIT6,4122)CMBID
CALL FETCH(8UOTS,CMBID,UDUR,USTART,UCNC,
*UUCNC,UFMSS,UUFMSS,UCMSS,UUCMSS,UTMSS,UUTMSS,MFLAG,
*DUM,ICNTRL,TYPE,UNIT6,UNIT 12,UNIT 13)
IBACK=2
IF (ICNTRL.E2-9)IRTN=1
IF(ICNTRL.E2.9)GO TO 500
CMB
IRTN=9
GO TO 2800
CONTINUE
CONTINUE
-INITIAL. INITIALIZE FITTING SOURCES
WRITE(UNIT6,3110)SIZNAM( SIZ)
DO 4410 IEL=1,NEL
ELFT(IEL,SIZ)=IBLANK
ELHOLD(IEL,SIZ)=0
DO 4420 JSOURC=1,NSOURC
SOUFT(JSOURC,SIZ)=IBLANK
SOHOLD(JSOURC,SIZ)=0
MEL=IMEL
MEL1(SIZ)=IMEL
MSOURC=IMSO
MSORC1(SIZ)=IMSO
DO 4450 KEL=1,MEL
IEL=INITEL(KEL)
AND ELEMENTS
136
-------
TABLE A.1 CMS MAIN PROGRAM (CONTINUED)
1585 ELFT(IEL,SIZ)=STAR
1586 ELHOLD(KEL,SIZ)=IEL
1587 4450 CONTINUE
1588 DO 4460 KSOURC=1,MSOURC
1589 JSOURC=INITSO(KSOURC)
1590 SOUFTCJSOURC,SIZ)=STAR
1591 SOHOLD(KSOURC,SIZ)=JSOURC
1592 4460 CONTINUE
1593 IFdRTN.ES. 0)GO TO 444
1594 GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
1595 C
1596 C
1597 C
1598 C
1599 C
1600 C -BACKOUT. SELECTS AND SUBTRACTS BACKGROUND
1601 4600 CONTINUE
1602 IFdBACK.NE. 2)GO TO 4670
1603 4602 WRITE(UNIT6,4605)
1604 4605 FORMAT(' ENTER BACKGROUND CMB SITE CODE= XXXXXXXXXXXX')
1605 READ(UNITS,4607,ERR=4602)BSITE
1606 4607 FORMAT(3A4)
1607 WRITE(UNIT6,4216)
1608 READ(UNIT5,4217) BYEAR
1609 WRITE(UNIT6,4219)
1610 READ(UNIT5,4220) BDATE
1611 ICNTRL=7
1612 4610 CALL FETCH(CUOTB,BACKID,BDUR,BSTART,BCNC,
1613 *UBCNC,BFMSS,UBFMSS,BCMSS,UBCMSS,BTMSS,UBTMSS,MFLAGB,
1614 *DUM,ICNTRL,TYPE,UNIT6,UNIT 12,UNIT 13)
1615 IFdCNTRL. E2. 9)IRTN=1
1616 IFdCNTRL.Efi. 9)GO TO 500
1617 DO 4650 1= 1 , 3
1618 DO 4650 IEL=1,NEL
1619 UCNC(IEL,I)=UCNC(IEL,I)-BCNC(IEL,I )
1620 UUCNC(IEL,I) = ((UUCNC(IEL,I)**2)+UBCNC(IEL,I)**2)** . 5
1621 4650 CONTINUE
1622 UFMSS=UFMSS-BFMSS
I623 UUFMSS=(UUFMSS**2+UBFMSS**2)**.5
1624 UCMSS=UCMSS-BCMSS
1625 UUCMSS=(UUCMSS**2+UBCMSS**2)**.5
1626 UTMSS=UTMSS-BTMSS
1627 UUTMSS=(UUTMSS**2+UBTMSS**2)**.5
1628 WRITE(UNIT6,4655)
1629 4655 FORMAT(' BACKOUT COMPLETE1)
1630 GO TO '4680
1631 4670 WRITE(UNIT6,4675)
1632 4675 FORMAT(* CURRENT DATA ALREADY HAS BACKGROUND SUBTRACTED')
137
-------
TABLE A.I CMS MAIN PROGRAM (CONTINUED)
1633 4680 IBACK=1
1634 GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
1635 C
1636 C
1637 C
1638 C -BACKIN. ELIMINATES CURRENT BACKGROUND SUBTRACTION
1639 4700 CONTINUE
1640 IFdBACK.NE. 1 )GO TO 4770
1641 DO 4750 1=1,3
1642 DO 4750 IEL=1,NEL
1643 UCNC(IEL,I)=UCNC(IEL,I)+BCNC(IEL,I)
1644 UUCNC(IEL,I)=C(UUCNC(IEL,I)**2)-UBCNC(IEL,I)**2)**.5
1645 4750 CONTINUE
1646 UFMSS=UFMSS+BFMSS
1647 UUFriSS=(UUFMSS**2-UBFMSS**2)**.5
1648 UCMSS=UCMSS+BCMSS
1649 UUCMSS=(UUCMSS**2-UBCMSS**2)**.5
1650 UTMSS=UTMSS+BTMSS
1651 UUTMSS=(UUTMSS**2-UBTMSS**2)**.5
1652 WRITE(UNIT6,4755)
1653 4755 FORMAK' BACKIN COMPLETE')
1654 GO TO 4780
1655 4770 WRITE(UNIT6,4775)
1656 4775 FORMATUH ,'CURRENT CMB DATA DOES NOT HAVE BACKGROUND SUBTRACTED')
1657 4780 IBACK=2
1658 GO TO (500,2860,4120,4130,4132,4150,4250,4260,4270,4280),IRTN
1659 C
1660 C
1661 C
1662 C
1663 C - PCOMP
1664 C
1665 C
1666 COMPUTE MEAN AND STANDARD DEVIATION OF SERIES
1667 52000 UNIT=UNIT11
1668 MRITE(UNIT6,52001)
1669 52001 FORMATC OUTPUT WILL GO TO HARDCOPY.'/
1670 * ' DO YOU WANT IT DISPLAYED AT YOUR TERMINAL INSTEAD?')
1671 READ(UNIT5,28001)ANS
1672 IF(ANS.EB.YESY)UNIT=UNIT6
1673 REWIND UNIT10
1674 DO 52005 I=1,MELMAX
1675 DO 52005 J=1,MSOMAX
1676 F(I,J)=0.0
1677 XP(I,J)=0.0
1678 52005 SF(I,J)=0.0
1679 K=0
1680 M=0
138
-------
TABLE A.1 CUB MAIN PROGRAM
(CONTINUED)
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
169H
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
17 1 1
17 12
1713
17 14
1715
1716
17 17
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
52010
5201 1
52020
52030
52032
52040
52050
52052
52055
52060
53000
53970
53972
READ(UNIT 10,52011,END = 53000)TEMPID,J,FINE,COARSE,TOTAL
FORMAT(3A4,3A2,1X,I2,11X,F8.3,8X,F8.3,8X,F8.3)
DO 52020 1=1,6
IF(TEMPID(I).NE.LASTID(I) .OR. K.EB.O)GO TO 52030
CONTINUE
GO TO 52040
K=K+1
IF(K.GT.16)GO TO 53000
DO 52032 1=1,6
LASTID(I)=TEMPID(I)
PNAM(K,I)=TEMPID(I)
CONTINUE
IF(M.GT.O)GO TO 52050
ITM(1)=J
11=1
IEL=1
GO TO 52060
DO 52052 1=1,M
IF(J.E2.ITM(I)) GO TO 52055
CONTINUE
IF(M.GE.21)GO TO 52010
ITM(M)=J
IEL = M
GO TO 52060
IEL = I
F(IEL,K)=FINE
XP(IEL,K)=COARSE
SF(IEL,K)=TOTAL
GO TO 52010
DO 53090 1=1,M
WRITECUNIT,53970)
FORMAT(/40X, 'FINE',9X, 'COARSE' ,7X, 'TOTAL' ,/' ' ,3X,
* 'CMB SITE1 ,7X, 'DATE1 ,4X, 'SOURCE1 ,6X, ' (UG/M3) ' ,6X,
* ' (UG/M3)' ,6X, '(UG/M3)')
TX = 0
UTX = 0
PCNT=0
UPCNT=0
WRITE(UNIT,53972)
FORMAT(4X,8('-'),7X,4('-'),4X,6('-'),6X,7('-'),6X,7('-'),
* 6X,7('-')/)
INDX=ITM(I)
TEMPN=SOUNAMCINDX)
DO 53080 J=1,K
WRITE (UNIT, 5 3 97 8) ( PNAM( vJ , IEL ) ,IEL=1 ,6) , TEMPN , F ( I, J ) ,XP(I, J) ,
139
-------
TABLE A.I CUB MAIN PROGRAM
(CONTINUED)
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
53978
53080
53988
53989
53990
53090
C
C
5000
C
C
C
: SFCI.J)
FORMAT(2X,3A4,3X,A2,'/I,A2,'/',A2,2X,A8,
: 2X,F7.2,7X,F7.2,6X,F7.2)
X=X+F(I,J)
TX=TX+XP(I,J)
PCNT=PCNT+SF(I,J)
IF(K.LE.1)GO TO 53080
UX=UX+F(I,J)**2
UTX=UTX+XP(I,J)**2
UPCNT=UPCNT+SF(I,J)**2
CONTINUE
WRITECUNIT,53988)
FORMAT(/39X,7(I-'),6X,7(f-l),6X,7(1-f))
IFCK.LE.1)GO TO 53989
UX = S6RT( (UX-X/K*X)/(K-1 . ) )
UTX = S2RT((UTX-TX/K*TX)/(K-1 . ) )
UPCNT=SfcRT((UPCNT-PCNT/K*PCNT)/(K-1.))
X = X/K
TX=TX/K
PCNT=PCNT/K
WRITE(UNIT,53990)X,TX,PCNT,UX,UTX,UPCNT
FORMAT(22X,'AVERAGE1,8X,F7.2,7X,F7.2,6X,F7.2/20X,
« '(STD. DEV.)',4X,'+-',F7.2,5X,'+-',F7.2,4X,'+-',F7.2/)
CONTINUE
REWIND UNIT10
IRTN=1
GO TO 500
STOP
END
140
-------
TABLE A.2 CUBE SUBROUTINE
1 SUBROUTINE CMBR(F,C,N,L,NN,LL,SF,V,VK,TOTALA,SSET,
2 * KSTARK,SAS,SAA,SR2,R2LS,VIF,
3 * RIDGE,EFVAR,INTCEP,IERR,KBUG,CWT,
4 * W,X,Y,XMEAN,XPX,XPXK,XPY>SX,S,TM,A,SOLD,IS,IA.ITM
5 * )
6 C THIS SUBPROGRAM PERFORMS THE CMB ANALYSIS
7 C OPTIONS ARE RIDGE REGRESSION,
8 C EFFECTIVE VARIANCE METHOD,
9 C AND INCLUSION OF INTERCEPT
10 C
1 1 C
12 C WRITTEN BY D.A. DUBOSE AND H.J. WILLIAMSON
13 C RADIAN CORPORATION,AUSTIN TEXAS IN 1982
14 C UNDER CONTRACT TO EPA (PROJECT OFFICER W.P. FREAS)
15 C
16 C
17 c ***********ARGUMENTS TO THE SUBPROGRAM**************************
18 C
ig c ***************** INPUT *********************
20 C
21 C F-SOURCE SIGNATURE MATRIX
22 C C-CONCENTRATION VECTOR
23 C SF-STANDARD ERROR OF F
24 C V-STANDARD ERROR OF C
25 C N-NUMBER OF SPECIES
26 C L-NUMBER OF SOURCES
27 C NN-MAXIMUM NUMBER OF SPECIES(ARRAY DIMENSION)
28 C LL-MAXIMUM NUMBER OF SOURCES(ARRAY DIMENSION)
29 C VK-VECTOR OF K-VALUES FOR RIDGE REGRESSION
30 C TOTALA-TOTAL AEROSOL
31 C
32 C
33 Q ***************** CONTROL*********************
34 C
35 C RIDGE-RIDGE OPTION FLAG 0=NO 1=YES 2=YES +SUMMARY
36 C EFVAR-EFFECTIVE VARIANCE FLAG 0=NO 1=YES
37 C INTCEP-INTERCEPT OPTION FLAG 0=NO 1=YES
38 C CWT-UEIGHT FOR RIDGE SOLUTION SELECTION
39 C KBUG-NOT USED IN THIS VERSION
40 C
41 C
42 C ***************** OUTPUT **********************
43 C
'4'4 C SSET-STANDARD ERROR OF CALCULATED TOTAL AEROSOL
45 C KSTARK-INDEX OF SELECTED RIDGE SOLUTION
46 C SAS-MATRIX OF RIDGE SOLUTIONS
47 C SAA-MATRIX OF STANDARD ERRORS OF SAS
48 C SR2-R-S2UARE OF SELECTED RIDGE SOLUTION
-------
TABLE A.2 CMBR SUBROUTINE (CONTINUED)
49 C R2LS-R-S2UARE OF LEAST SQUARES SOLUTION
50 C VIF-VARIANCE INFLATION FACTORS
51 C IERR-INVERSION ERROR FLAG
52 C
53 C
54 c ********************* WORKSPACE ********************
55 C
56 C W-WEIGHTS
57 C X-SCALED SIGNATURE
58 C Y-SCALED CONCENTRATIONS
59 C XMEAN-SOURCE MEANS
60 C XPX-X'X MATRIX (CORRELATION MATRIX)
61 C XPXK-XPX INVERSE
62 C XPY-X'Y VECTOR
63 C SX-SOURCE SCALE FACTORS
64 C S-SOLUTION VECTOR
65 C TM-MATRIX MULTIPLICATION WORK SPACE
66 C A-STANDARD ERROR OF SOLUTION VECTOR S
67 C SOLD -HOLDER FOR SOLUTION TO CHECK FOR CONVERGENCE
68 C IA-WORKSPACE FOR INVERSION ROUTINE
69 C ITM -WORKSPACE FOR INVERSION ROUTINE
70 C IS -WORKSPACE FOR INVERSION ROUTINE
71 C
72 C
73 C
714 c ******** OTHER VARIBLES AND ARRAYS*************
75 . C
76 C ADIF-RIDGE SELECTION CLOSENESS CRITERIA VALUE
77 C AINT-INTERCEPT
78 C DET-ARGUMENT FOR INVERSION ROUTINE(DETERMINENT)
79 C DF-DEGREES OF FREEDOM
80 C EPS-ARGUMENT FOR INVERSION ROUTINE
81 C IAGAIN-REITERATION FLAG
82 C IBLANK-BLANK ' '
83 C IN-NUMBER OF NEGATIVE SOLUTIONS
84 C ITER-ITERATION NUMBER
85 C KALL-NUMBER OF RIDGE K-VALUES
86 C KINDEX - INDEX OF RIDGE SOLUTION
87 C LP1-L PLUS ONE FOR INTERCEPT IF RESUESTED
88 C MAXIT-MAXIMUM NUMBER OF ITERATIONS
89 C MORE-NUMBER OF ADDITIONAL ITERATIONS
90 C MSE-MEAN SfiUARE ERROR
91 C NPIV.PIV-ARGUMENTS FOR INVERSE ROUTINE
92 C SE-STANDARS ERROR(S2RT USE)
93 C SET-STANDARD ERROR OF CALCULATED TOTAL AEROSOL
94 C SSE-SUM OF SQUARES FOR ERROR
95 C SY-CONCENTRATION SCALE FACTOR
96 C SYSX-SCALE FACTOR FOR COEFFICIENTS
142
-------
TABLE A.2 CMBR SUBROUTINE (CONTINUED)
97 C UNITS,UNITS-FORTRAN READ/WRITE UNITS
98 C WORST-VALUE OF NEGATIVE COEFFICIENTS OF GREATEST MAGNITUDE
99 C WT-WEIGHTCOFTEN USED AS WORK VARIABLE)
100 C STARK-K-VALUE OF BEST RIDGE SOLUTION TO DATE
101 C SUMNEG-NEGATIVE COEFFICIENT RIDGE SELECTION CRITERIA VALUE
102 C SUMWT-RIDGE SELECTION CLOSENESS(ALSO WORK SPACE)
103 C SXM-WORK VARIABLE
104 C
105 C *****************************************************************
106 REAL F(NN,LL), C(NN), V(NN), SFCNN,LL), VKC31),
107 * XCNN,LL),SAS(LL,31),SAA(LL,31),VIF(LL),
108 * Y(NN),XMEAN(LL),XPX(LL,LL),XPXK(LL,LL),
109 * XPY(LL), SX(LL), S(LL), TM(NN), A(LL), SOLD(LL),W(NN),SSET(2)
110 INTEGER IS(LL),IA(LL),ITM(NN),UNITS,UNIT6
1 1 1 REAL K,MSE
112 INTEGER EFVAR,RIDGE
113 DATA IBLANK/' '/
1 14 C
115 UNIT5=5
116 UNIT6=6
117 KALL=31
1 18 C
119 EPS=0.
120 ADIFF=1.E30
121 KINDEX=1
122 DO 2020 1=1,N
123 W(I)=V(I)
124 2020 CONTINUE
125 KBUG=0
126 CONTROL POINT - BEGIN CYCLE FOR K-VALUE
127 3000 CONTINUE
128 K=VK(KINDEX)
129 MAXIT=10
130 ITER=1
131 IFCEFVAR.ES.0 .AND. KINDEX.GT.1) GO TO 9000
132 CONTROL POINT - BEGIN ITERATION LOOP
133 4000 CONTINUE
134 COMPUTE WEIGHTS £ WEIGHTED MEANS
135 SUMWT=0.
136 YMEAN=0.
137 DO 5005 J=1,L
138 5005 XMEANCJ)=0.
139 DO 5008 1=1,N
140 W(I)=1./W(I)
I'M 5008 CONTINUE
142 LP1=L
143 IFCINTCEP.EQ.0) GO TO 6000
144 CORRECT SOURCE MATRIX TO INCLUDE INTERCEPT TERM
143
-------
TABLE A.2 CMBR SUBROUTINE
(CONTINUED)
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
19 1
1 92
5020
6000
CALCUI
6010
6020
COMPU1
6050
6060
6070
COMPU'
6080
6090
COMPU'
7010
7020
7025
7030
L P 1 -•- L + 1
DO 5020 1=1,N
FCI.LP1) = 1 .0
CONTINUE
CONTINUE
,ATE VARIANCE STABILIZATION BY WEIGHTING
DO 6020 1=1,N
DO 6010 J=1,LP1
X(I,J) = F(I,J)*M(I)
CONTINUE
Y(I)=C(I)*W(I)
CONTINUE
:E STANDARDIZING FACTORS
DO 6060 J=1,LP1
WT = 0 .
DO 6050 1=1,N
WT=WT+X(I,J)**2
CONTINUE
SXCJ)=SfiRT(WT)
CONTINUE
SY = 0 .
DO 6070 1=1,N
SY=SY+Y(I)**2
CONTINUE
SY=S2RT(SY)
CE STANDARDIZED DATA VALUES
DO 6090 1=1,N
DO 6080 J=1,LP1
X(I,J)=X(I,J)/SX(J)
CONTINUE
Y(I)=Y(I)/SY
CONTINUE
FE MATRIX OF SUM OF SQUARES AND CROSS PRODUCTS
DO 7030 J=1,LP1
DO 7020 11=1 , J
WT = 0
DO 7010 1=1,N
WT=WT + X(I,M) * X(I.J)
CONTINUE
XPX(M,J) = WT
XPX(J,M) = WT
CONTINUE
WT = 0
DO 7025 1=1,N
WT=WT + X(I,J)*Y(I)
CONTINUE
XPY(J)=WT
CONTINUE
144
-------
TABLE A.2 CMBR SUBROUTINE (CONTINUED)
193 CHECKPOINT - ADD RIDGE TO SSCP MATRIX
194 9000 CONTINUE
195 DO 9050 M=1,LP1
196 I=M-1
197 DO 9040 J=1,I
198 XPXKCM,J)=XPX(M,J)
199 XPXKCJ,M)=XPX(M,J)
200 9040 CONTINUE
201 XPX(M,M)=1.0 + K
202 XPXK(M,M)=XPX(M,M)
203 9050 CONTINUE
204 CALL INVERSE ROUTINE
205 CALL INV1(XPXK.LP1,LL,EPS,IS,IERR,DET,NPIV,PIV,IA,ITM)
206 IFdERR.EC. 0)GO TO 10000
207 WRITE(UNIT6,9090)
208 9090 FORMATC' ***ERROR= SINGULAR MATRIX, CHECK DATA')
209 IERR=1
210 GO TO 99900
211 COMPUTE SOLUTION VECTOR
212 10000 CONTINUE
213 DO 10020 J=1,LP1
214 WT=0.
2 15 DO 10010 11=1,LP1
216 WT = UT + XPXKCJ,M)*XPY(M)
217 10010 CONTINUE
218 S(J)=WT
219 10020 CONTINUE
220 COMPUTE R-S2UARED
221 SSE=0.
222 DO 12020 1=1,N
223 WT = 0.
224 DO 12010 J=1,LP1
225 WT=WT + X(I,J)*S(J)
226 12010 CONTINUE
227 WT=Y(I)-WT
228 SSE=SSE+WT*UT
229 12020 CONTINUE
230 WT=0.
23 1 IF(INTCEP.E6. 1 )WT = XPY(LP1)**2
232 R2=1-SSE/(1-WT)
233 DF=N-LP1
234 MSE=SSE/DF
235 SE=S2RT(MSE)
236 COMPUTE STANDARD ERRORS
237 SET=0.0
238 DO 14030 M=1,LP1
239 DO 14020 J=1,LP1
240 WT = 0
145
-------
TABLE A.2 CMBR SUBROUTINE
(CONTINUED)
241 DO 14010 1=1,LP1
242 WT=WT + XPXK(M,I)*XPX(I,J)
243 14010 CONTINUE
244 TM(J)=WT - K*XPXK(M,J)
245 14020 CONTINUE
246 WT=0
247 DO 14025 J=1,LP1
248 WT = WT + TM(J)*XPXK(J,M)
249 COMPUTE DATA SCALE STANDARD ERROR
250 SXM=SX(M)
251 DO 14025 1=1,LP1
252 SET=SET+TM(J)/SXM*XPXK(J,I)/SX(I)
253 14025 CONTINUE
254 A(M)=SE*S6RT(WT)
255 14030 CONTINUE
256 SET=SBRT(SY*SET)
257 COMPUTE DATA SCALE SOLUTION AND STANDARD
258 AINT=0.0
259 DO 16020 J=1,LP1
260 SYSX=SY/SX(J)
261 SCJ)=SCJ)*SYSX
262 ACJ)=A(J)*SYSX
263 16020 CONTINUE
264 IFCINTCEP.NE.0)AINT = S(LP1 )
265 COMPUTE EFFECTIVE VARIANCES
266 IF(EFVAR.ES.O) GO TO 20000
267 DO 17030 1=1,N
268 WT=V(I)**2
269 DO 17020 J=1,L
270 SYSX=S(J)*SF(I,J)
271 WT=UT + SYSX*SYSX
272 17020 CONTINUE
273 W(I)=S2RT(WT)
274 17030 CONTINUE
275 CHECK FOR CONVERGENCE
276 IAGAIN=0
277 IF(ITER.E2.1) GO TO 18050
278 DO 18030 J=1,LP1
279 WT=ABS(SCJ)-SOLD(J))/A(J)
280 IFCWT.GT. 0.1) IAGAIN=1
281 18030 CONTINUE
282 IFCIAGAIN.E2.0 .OR. ITER.GE.MAXIT)
283 18050 ITER = ITER-H
284 DO 18070 J=1,LP1
285 SOLD(J)=S(J)
286 18070 CONTINUE
287 CIRCLE BACK FOR ANOTHER ITERATION
288 GO TO 4000
OF TOTAL AEROSOL
ERRORS
GO TO 20000
146
-------
TABLE A.2 CMBR SUBROUTINE (CONTINUED)
289 COMPLETED ITERATION FOR THIS K
290 20000 CONTINUE
291 SSE=SSE*SY*SY
292 COMPUTE TOTAL AEROSOL AND t NEGATIVE COEFFICIENTS
293 WORST=0.
294 SUMNEG=0.
295 IN=0
296 SUMWT=0.
297 DO 20080 J=1,L
298 IF(S(J).GE.0.)GO TO 20078
299 IN=IN+1
300 SUMNEG=SUMNEG+SCJ)*SCJ)
301 IF(S(J) .LT.WORST) WORST = S(J)
302 20078 SUMWT=SUMWT+SCJ)
303 20080 CONTINUE
304 SUMWT=SUMWT+AINT
305 IF(RIDGE.NE.2) GO TO 21099
306 IFCKINDEX.E2.1)WRITE(UNIT6,21085)
307 21085 FORMATC/1 RIDGE K',2X,'R-S2UARE',3X,'AEROSOL',3X,'SE(AEROSOL)
308 * 2X,'# NEC',MX,'WORST1,2X, 'ITERATIONS')
309 WRITE(UNIT6,21086)K,R2,SUMWT,SET,IN,WORST,ITER
310 21086 FORMATC2X,F5.3,4X,F6.4,3X,F8.3,3X,F8.2,6X,I2,F11.3,5X,
311 * 13)
312 21099 CONTINUE
313 IFCITER.LT.MAXIT .OR. IAGAIN.Eft.0)GO TO 28000
314 22149 IF(RIDGE.GT.O)WRITE(UNIT6,22150)ITER,K
315 22150 FORMATC' ',13,' ITERATIONS SO FAR FOR RIDGE K=',F5.3,
316 * ' WITHOUT CONVERGENCE'/' HOW MANY MORE DO YOU WANT TO ',
317 * 'TRY?')
318 IFCRIDGE.E2.0)WRITECUNIT6,22151)ITER
319 22151 FORMATC' ',13,' ITERATIONS SO FAR WITHOUT CONVERGENCE'/
320 * ' HOW MANY MORE DO YOU WANT TO TRY?')
321 READC UNITS,22152,ERR=22149)MORE,I
322 22152 FORMAT(12,T2,A 1 )
323 IFCI.E8.IBLANK)MORE=MORE/10
324 ITER=ITER+1
325 MAXIT=MAXIT+MORE
326 IF CMAXIT.GT.ITER)GO TO 4000
327 28000 CONTINUE
328 CHOCK AWAY COEFFICIENTS, ETC. FOR THIS RIDGE K SOLUTION
329 IFCKINDEX.E2.1)SSETC1)=SET
330 DO 27030 J=1,LP1
331 SASCJ,KINDEX)=SCJ)
332 SAACJ,KINDEX)=ACJ)
333 27030 CONTINUE
334 COMPUTE VARIANCE INFLATION FACTORS
335 IFCK.NE.0.0) GO TO 27050
336 R2LS=R2
147
-------
TABLE A.2 CMBR SUBROUTINE (CONTINUED)
337 DO 27040 J=1,L
338 VIF(J)=XPXK(J,J)
339 27040 CONTINUE
340 27050 CONTINUE
341 CHOOSE BETTER SOLUTION
342 SUMUT=(SUMWT-TOTALA)**2 + CWT*SUMNEG
343 IFCSUMWT.GT.ADIFF) GO TO 40000
344 ADIFF=SUMWT
345 STARK=K
346 KSTARK=KINDEX
347 SR2=R2
348 SSET(2)=SET
349 CHECK FOR LAST OF K-SOLUTIONS
350 40000 CONTINUE
351 IFCKALL.LE.KINDEX .OR. RIDGE.Efi.O) GO TO 50000
352 KINDEX = KINDEX-M
353 CIRCLE BACK FOR ANOTHER K-VALUE
354 GO TO 3000
355 COMPLETED ALL K-VALUES
356 50000 CONTINUE
357 IFCRIDGE.GE.2)WRITE(UNIT6,50010) STARK
358 50010 FORMATC'ORIDGE K= ',F5.3,' IS THE SELECTED BEST SOLUTION')
359 99900 RETURN
360 END
148
-------
TABLE A.3 FETCH SUBROUTINE
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
4 1
42
43
44
45
46
47
48
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
FETCH - ACCESSES INPUT DATA STORAGE
MODIFIED 1982 BY D.A. DUBOSE
UNDER CONTRACT TO US EPA
OF RADIAN CORP,AUSTIN TX
ICNTRLS:
0- AUTOFIT EOF
1- AUTOFIT START
2- AUTOFIT IN MID SEARCH
4- SELECT START
5- SELECT IN MID SEARCH
7- BACKGROUND START
8- BACKGROUND IN MID SEARCH
9- DATA SEARCH FAILED
SUBROUTINE FETCH(ASB,INID,DUR,START,
*CONC,UCONC,FMASS,UFMASS,CMASS,UCMASS,TMASS,UTMASS,
*MFLAG,SAVE,ICNTRL,TYPE,UNIT6,UNIT12,UNIT13)
REAL*4 FMASS,UFMASS,CMASS,UCMASS,TMASS,UTMASS,CONG(35,3)
*UCONC(35,3),
*T10,T7,T8,T9
CHARACTER VARIABLES
INTEGER*4 ASB , SAVE ( 5 ), TYPE
* ,MFLAG(35, 3) ,
* eUOTM,BLANK,2UOTB,eUOTA,eUOTS,2UOTFT,eUOTFC
INTEGER SPECS
INTEGER* 4 TO , T6 , SCODE ( 35 ) ,PCODE( 35) , DUR, START,
* UNIT6,UNIT12,UNIT13
CHARACTER VARIABLE SPECS
INTEGER*4 INID ( 5 ) , ID ( 5 )
COMMON /F/ PCODE, SCODE, NEL
EQUIVALENCE (SITE, INI D( 1 ) ) , ( YEAR , INID ( 4 ) ) , ( DATE , INID ( 5 ) )
DATA INITIALIZATION
DATA BUOTM, BLANK, fiUOTB
* 'M',' ' , 'B1 , 'A' , 'S'
BUOTA , 2UOTS
'FT' , 'FC'/
2UOTFT , CUOTFC /
149
-------
TABLE A.3 FETCH SUBROUTINE (CONTINUED)
49 C
50 C
51 C
52 C
53 C
54 C FORMAT STATEMENTS
55 C TITLE CARD FOR SITE NUMBER AND YEAR AND DATE
56 25 FORMAT(5X,' SITE: ',3A4,5X,'YEAR: ',A2,2X,'DATE= ',A4,5X,A2)
57 C POLLUTANT DATA CARDS
58 155 FORMAKI2, 1X,3A4, 1X.A2.A4, 1X,I2, 1 X , 12,4X,12,2X,
59 *F9.4,2X,F9.4,2X,F9.<4,2X,F9.4)
60 C
61 C
62 C DATA SEARCH FAILURE
63 410 FORMATdH , ' DATA SET NOT FOUND FOR' /6X, ' SITE : ',
64 * 3A4,5X,'YEAR: ',A2,2X,'DATE: ',A4/
65 * ' USE ONE OF THOSE LISTED ABOVE')
66 C
67 C
68 C
69 C
70 C
71 C INITIALIZE CONCS,MASSES AND MISSING DATA FLAGS
72 DO 3 1=1,NEL
73 DO 3 J=1,3
74 CONCCZ, J)=0\000
75 UCONCd, J)=0.001
76 MFLAGd, J)=fiUOTM
77 3 CONTINUE
78 FMASS=0.000
79 UFMASS=0.000
80 CMASS=0.000
81 UCMASS=0.000
82 TMASS=0.000
83 UTMASS=0.000
84 C
85 C
86 C
87 C
88 C
89 C
90 C
91 C REWIND DATA FILE
92 4 REWIND UNIT13
93 C
94 C
95 C
96 C
150
-------
TABLE A.3 FETCH SUBROUTINE (CONTINUED)
97 C READ DATA LINE
98 10 READ (UNIT13,155,END=200)TO,ID,IDUR,ISTART,T6,T7,T8,T9,T10
99 C
100 C
101 C
102 C
103 C
104 C
105 C
106 C
107 C
108 C ROUTE CARD TYPES
109 IF(TO.E2.3)GO TO 30
110 IF(TO.E2.30.AND.ICNTRL.E2.2)GO TO 150
111 IF(TO.E2.30.AND.ICNTRL.E2.5)GO TO 150
112 IFCTO.E2.30.AND.ICNTRL.EC.8)GO TO 150
113 GO TO 10
114 C
1 15 C
1 16 C
1 17 C
118 CONTROL POINT - CARD TYPE 3 (SITE HEADER CARD)
119 30 CONTINUE
120 DO 35 1=1,5
121 SAVE(I)=ID(I)
122 35 CONTINUE
123 C CHECK CONTROLS
124 IFdCNTRL.EB. 2)GO TO 320
125 IF(ICNTRL.E2.8)GO TO 320
126 IF(ICNTRL.E2.5)GO TO 320
127 TYPE=2UOTFC
128 IF(T6.E2.13)TYPE=2UOTFT
129 WRITE(UNIT6,25)ID,TYPE
130 C MATCH INID
131 DO 37 1=1 ,5
132 IF(IDd) .NE.INID(I) )GO TO 10
133 37 CONTINUE
134 DUR=IDUR
135 START=ISTART
136 IF(ASB.NE.2UOTB)WRITE(UNIT12,155) TO,ID,DUR,START,T6
137 C ADJUST CONTROLS
138 IF(ASB.E2.2UOTA)ICNTRL=2
139 IF(ASB.E2.2UOTB)ICNTRL=8
140 IF(ASB.E2.2UOTS)ICNTRL=5
141 C BACK TO READ
142 GO TO 10
143 C
144 C
151
-------
TABLE A.3 FETCH SUBROUTIKE
(CONTINUED)
1<45
146
147
143
149
150
151
152
153
154
155
156
157
153
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
C
C
C
C
C
C
C 1
150
C (
C
165
167
C M
168
169
170
POLLUTANT CARDS
CONTINUE
CHECK IF POLLUTANT IS MASS
IFd6.NE.01)GO TO 168
IF(TYPE .NE. 2UOTFT)GO TO 167
FMASS=T7
UFMASS=T8
TMASS=T9
UTHASS=T10
CALL COARS(CMASS,UCMASS,TMASS,UTMASS,FMASS,UF11ASS)
GO TO 168
FMASS=T7
UFMASS=T8
CMASS=T9
UCMASS=T10
TMASS=CMASS+FMASS
UTMASS=SeRT(UCMASS*UCMASS+UFMASS*UFMASS)
GO TO 168
MATCH POLLUTANT CONCENTRATIONS
DO 172 1=1,NEL
IF(T6 .ES.PCODEd) )GO TO 169
GO TO 172
IFCTYPE .NE. BUOTFT)GO TO 170
CONCd,1)=T7
UCONCd,1)=T8
CONC(I,3)=T9
UCONCd,3)=T10
IFCCONCd, 1).E2.0
IF(CONC(I,3) .E2.0
CALL COARSCCONCCI
*CONC(I, 1 ) ,UCONC(I
GO TO 171
CONCCI,1)=T7
UCONCCI,1)=T8
CONCCI,2)=T9
UCONCCI,2)=T10
CONC(I,3)=CONC(I,1)+CONC(I,2)
UCONCd, 3 )=S2RT( UCONCCI, 1 ) **2 + UCONC (1, 2 ) **2 )
IF (CONG (I, 1 ) .E2. 0 . 000 . AND. UCONCd, 1 ) .EB. 0 . 000) UCONCd
IF (CONG (1,2) . ES. 0.000 . AND. UCONCd, 2 ) . E2 - 0 . 000) UCONCd
000. AND. UCONCd, 1 ) .E8. 0 . 000) UCONCd, 1 ) =000 .001
000. AND. UCONCd, 3) .Efi. 0 . 000) UCONCCI, 3) =000 . 00 1
2 ) , UCONCd, 2) ,CONC(I, 3) , UCONCd, 3) ,
1 ) )
1)=000
2)=000
IF (CONG (I, 3) . E6. 0.000 .AND. UCONCd, 3) .E2. 0 . 00 0 ) UCONC (1, 3) =000
001
001
001
152
-------
TABLE A.3 FETCH SUBROUTINE
(CONTINUED)
1 93
19'4
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
21 1
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
C ADJUST MISSING DATA FLAGS
171 MFLAGCI,1)=BLANK
MFLAGCI,2)=BLANK
MFLAGCI,3)=BLANK
CONTINUE
SELECT OR AUTOFIT,WRITE TEMP
172
C IF
175
j^ jj j-i w x v/x\. nw.L\si.j-.i.yi*f£\..*.j.jrf j.j-111* ji
IFCASB.NE.eUOTB)WRITE(UNIT12,155)TO,INID,DUR,START,T6,T7, T8,T9,
* T10
ADJUST CONTROLS
IFCASB.E2.2UOTA)ICNTRL=2
IF(ASB.EB.2UOTB)ICNTRL=8
IFCASB.Eg.eUOTS)ICNTRL=5
BACK TO READ
GO TO 10
C
C
C
C
C
C
C
200
END OF FILE
IFCICNTRL.E6
IFCICNTRL.Efi
IFdCNTRL.EB
. 1 )GO
,7)GO
.4)GO
TO
TO
TO
400
400
400
IF(ICNTRL.E2.2)ICNTRL=0
GO TO 320
C
C
C
C
C
C
C RETURN TO MAIN
320 RETURN
C
C
C DATA SEARCH FAILED
400 WRITE(UNIT6,410)INID
ICNTRL=9
RETURN
END
153
-------
TABLE A.H COARS SUBROUTINE
1 C SUBROUTINE COARS
2 C FOR PROGRAM CMB
3 C BY DR. J.G.WATSON
4 C GETS COARSE FRACTION AND UNCERTAINTY
5 C
6 SUBROUTINE COARS(X,UX,TCONC,UTCONC,FCONC,UFCONC)
7 IF(UTCONC.E2.-1..OR.UFCONC.E2.-1.)GO TO 10
8 X=TCONC-FCONC
9 UX=(UTCONC**2+UFCONC**2)**.5
10 IF(X.LT.UX)X=UX/2.
1 1 RETURN
12 10 X=-1.
13 UX=-1.
14 RETURN
15 END
154
-------
TABLE A.5 PERC SUBROUTINE
1 C
2 C
3 C
4 C SUBROUTINE PERC
5 C FOR PROGRAM CMB
6 C BY D. W. TORKELSON
7 C THIS ROUTINE CALCULATES PERCENTS AND THEIR UNCERTAINTIES
8 C
9 SUBROUTINE PERCCPCNT,UPCNT,X1,UX1,X2,UX2)
10 REAL*4 PCNT,UPCNT,X1,UX1,X2,UX2
) 1 IFCX2.EZ.0. )GO TO 5
'2 IFCX1 .EB.-1 . .OR.X2.E2.-1.)GO TO 15
13 PCNT=100.*(X1/X2)
!4 UPCNT=100.*((UX1/X2)**2+(UX2*X1XX2**2)**2)**.5
1 5 IF(PCNT.LT.UPCNT)PCNT = UPCNT/2.
16 GO TO 10
17 5 PCNT=0.
18 UPCNT=0.
19 10 RETURN
20 15 PCNT=-1.
21 UPCNT=-1.
22 RETURN
2 3 END
24 C
155
-------
TABLE A.6 INV1 SUBROUTINE
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
I N ',
I N V -
IN"
INVO.
INVO
BE HELPFUL IN FOLLOWING THE CODE.
THE CALLING PROGRAM MUST SET A , N, NN, EPS , AND LTEMP TO--
SUBROUTINE INV1(A,N,NN,EPS,LTEMP,IERR,DET,NPIV,PIV,LPR,LPC) INVO<
DIMENSION A(NN,NN) INVC
DIMENSION LTEMP(1),LPR(1),LPC(1) INVO
DECK 8045A
SUBROUTINES CALLED - NONE
THIS ROUTINE INVERTS MATRIX A IN ITS OWN SPACE. IT ALSO COMPUTES THE
THE DETERMINANT OF A.
THE METHOD IS THE USUAL GAUSSIAN EXCHANGE PROCESS. BOTH ROWS AND INVO
COLUMNS ARE SEARCHED FOR MAXIMAL PIVOTS. THERE IS NO UNNECESSARY INVO
INTERCHANGING OF ROWS OR COLUMNS, ALL SUCH INTERCHANGES BEING CARRIED INVC
OUT AFTER THE EXCHANGE PROCESS IS COMPLETE. CHAPTER 1 OF E.L. STIEFLE,INVO
INTRODUCTION TO NUMERICAL MATHEMATICS,ACADEMIC PRESS,N.Y.,1963,SHOULD INVO
INVO
INVO
INVO
INVC
A-THE MATRIX TO BE INVERTED INVU
INVO;
INV
INVn.
INVO:
INVC
INVO.
INVO:
iNvr
INVu.
INVO
IN .'';
INVO
INVC
IK V
INV
INVO
1HV
INv,
INVO'
INV
INVt;
INVO'
INV;
INVO
INVO'
INV
LPR-THE FIRST NPIV POSITIONS LIST THE PIVOT ROW INDICES IN ORDER INV'
N-THE ORDER OF A
NN-THE NUMBER OF WORDS OF STORAGE PROVIDED FOR EACH COLUMN
OF ARRAY A BY THE CALLING PROGRAM
EPS-A NON-NEGATIVE NUMBER WHICH EACH PIVOT IS RE2UIRED TO EXCEED
IN ABSOLUTE VALUE (CUSTOMARILY ZERO)
LTEMP-A BLOCK OF AT LEAST N WORDS OF TEMPORARY INTEGER STORAGE
IN ADDITION TO OVERWRITING A WITH ITS INVERSE,THE ROUTINE ALSO SETS
IERR,DET,NPIV,PIV,LPR,AND LPC TO-
IERR- 0 IF INVERSION IS COMPLETED AND NO TROUBLE IS DETECTED
2 IF MAGNITUDE OF CURRENT PIVOT FAILS TO EXCEED EPS
(INVERSION WILL NOT BE COMPLETED)
DET-PLUS OR MINUS THE PRODUCT OF THE CURRENT AND ALL PREVIOUS
PIVOTS
NPIV-THE NUMBER OF THE CURRENT PIVOT (FIRST,SECOND,ETC.)
PIV-THE CURRENT PIVOT
156
-------
TABLE A.6 INV1 SUBROUTINE
(CONTINUED)
49
50
5 1
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
7 1
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
OF USE
LPC-THE FIRST NPIV POSITIONS LIST THE PIVOT COLUMN INDICES IN
ORDER OF USE
DO INITIALIZATIONS
IERR=0
DET= 1 .
DO 2 1=1, N
LPR(I)=I
2 LPC(I)=I
BEGIN EXCHANGE PROCESS
DO 17 NP=1,N
NPIV=NP
SELECT PIVOT
PIV=0.
DO 4 K=NP,N
I=LPR(K)
DO 4 L=NP,N
J=LPC(L)
IF (ABS(A(I,J))-ABS(PIV)) 4,3,3
3 KPIV=K
INVO'
INVO
INV O.r
IPIV=I
JPIV=J
PIV=A(I, J)
4 CONTINUE
UPDATE DETERMINANT AND PIVOT ROM AND COLUMN LISTS
DET=DET*PIV
ITEMP=LPR(NP)
LPR(NP)=LPR(KPIV)
LPR(KPIV)=ITEMP
ITEMP=LPC(NP)
LPC(NP)=LPC(LPIV)
LPC(LPIV)=ITEMP
EXIT IF PIVOT TOO SMALL
IF (EPS-ABS(PIV)) 8,7,7
7 IERR=2
INVO
INVO1
INV(
INVO
INVO'
INV •'•
INVu
INVO'
INV
INVV
INVOf
INVOi
INVO
INVOf
INV'l
INV;'
INVO'
I?! .
INVO ,
INVO
INVO"
INVO'
INVC ,
INVO:
INVO '
INVO
INVO?
INVO'
INVC
INVOf
INV"
I N V 0 :
INVO;
INV '
INVO
INVOC
INV"-
INVO!
INVO<;
INVO'
INVO.
157
-------
TABLE A.6 IKV1 SUBROUTINE
(CONTINUED)
97
98
99
100
101
102
103
104
105
106
107
108
109
1 10
1 1 1
1 12
1 13
! 14
1 15
1 16
1 17
1 18
1 19
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
RETURN
MODIFY PIVOT ROW
8 DO 9 J=1,N
9 AdPIV, J)=-A(IPIV, J)/PIV
MODIFY OTHER ROWS
DO 14 1=1 ,N
IFCI-IPIV) 10,14, 10
10 TEMP=A(I, JPIV)
IF(TEMP)11 ,14,11
11 DO 13 J=1,N
IFCJ-JPIV) 12,13, 12
12 Ad, J)=A(I, J)+ACIPIV, J)*TEMP
13 CONTINUE
14 CONTINUE
MODIFY PIVOT COLUMN
DO 15 1=1 ,N
15 ACl,JPIV)=Ad,JPIV)/PIV
AdPIV, JPIV) =- A(IPIV, JPIV)
17 CONTINUE
END EXCHANGE PROCESS
UNSCRAMBLE ROWS OF INVERSE AND ADJUST SIGN OF DETERMINANT
DO 18 1=1, N
L=LPR(I)
18 LTEMP(L)=LPC(I)
DO 22 1=1, N
19 K=LTEMP(I)
IFd-K)20,22,20
20 DET=-DET
DO 21 J=1 ,N
TEMP=A(I, J)
A(I, J)=A(K, J)
21 A(K,J)=TEMP
LTEMP(I)=LTEMP(K)
LTEMP(K)=K
GO TO 19
22 CONTINUE
UNSCRAMBLE COLUMNS OF INVERSE
INVO'
INVO
INVO'
INV 1r
IN1' i
IKV I'
INV If
INV 1 '
INVH
INV I1'
IN',' 1
INV U
INV 1C
INV 1 '
INV1 ,
INV1 1
INV •
INV \
INV 1 i
INV1
INV 1
INV1 '
INV-
INV i L
INV1J
INV 1;
INV 1
INV i;
IN VI '
INV 1 -
INVI;
IKV i
INV 1.
INV 1 '
INV 1
INV1-
INV1 '
INV '
INV 1
INV1 "
INV i
INVI
INV1 .
INV 1 '
IKV i
INVV
INV ;
INV 1 '
158
-------
TABLE A.6 INV1 SUBROUTINE
(CONTINUED)
145 DO 23 1=1,N
146 L=LPC(I)
147 23 LTEMP(L)=LPR(I)
148 DO 27 1=1,N
149 24 K=LTEMP(I)
150 IF(I-K)25,27,25
151 25 DO 26 J= 1 ,N
152 TEMP=A(J,I)
153 A(J,I)=A(J,K)
154 26 A(J,K)=TEMP
155 LTEMP(I)=LTEMP(K)
156 LTEMP(K)=K
157 GO TO 24
158 27 CONTINUE
159 RETURN
160 END
INV 1 '
INV >
INV 1
INV 1!
INV '
INV 1
INV1:
INV1 '
INV1.'
INV1F
IN" •
INV ,
INV1.'
INV 1
INV 1
INV1f
159
-------
APPENDIX B
SELECTED FLOW DIAGRAMS
The diagrams presented in this appendix show the relationships among
commands, input/output units and subprograms. The diagrams do not give
specific details but overall concepts instead.
160
-------
I /
I /
/ I
AUTOFIT
*
RESUME
*Dashed lines indicate
Autofic control pach
COMMAND
LINK
Figure B-l. Generalized Program Flow Diagram
161
-------
COMMAND
LINK
WRITE
SOLUTION &
RESULTS
COMPUTE
TOTAL
WRITE
SUMMARY
WRITE
TEMPORARY
FILE
WRITE USER
FOLLOWUP
FILE
Figure B-2. WRITE Command Flow
162
-------
NEXT
RECEPTOR
COMMAND
LINK
AUTOFIT
(SELECT)
NY
CMB
(FINE)
NY
PDATA
(FINE)
WRITE
(FINE)
CMB
(COARSE)
PDATA
(COARSE)
VHIITE
(COARSE)
Figure B-3. AUTOFIT Command Flow
163
-------
GENERAL
HARDCOPY
TERMINAL
Figure B-4. General Hardcopy Communications
164
-------
MAIN
CMB
SELECT
AUTOFIT
WRITE
CMBR
Figure B-5.
Relationship of Main Program Commands and
Subprogram Linkages
165
-------
CENTER DATA'
STABILIZE VARIANCES
STANDARDIZE DATA
COMPUTE SSCP2 MATRIX
ADD fel co SSCP3
INVERT SSCP
COtlPUTE SOLUTION
COMPUTE STD. ERRORS
RESCALE SOLUTION
COMPUTE EFF. VAR. u
N0 / CON-
VERGENT*
YES
'If intercept is called for.
2SSCP denotes sum of squares and cross produces.
:I is the identity matrix.
'These items apply if the option to use effective variances is chosen.
'If ridge regression is not selected, 0 is the only value of fe to be used.
Figure B-6 . General CMBR Program Flow Diagram
166
-------
APPENDIX C
INPUT AND OUTPUT FOR EXAMPLE RUN
This appendix contains example input and output file listings for the
CMB program. The data given in Tables C-l through C-5 are the input files
used for the example interactive session of Section 4. The source signature
data presented in Tables C-3 and C-4 were compiled by Dzubay and Hasan
(1983). Accurate estimates of the standard errors are not currently avail-
able for all of the source signatures. A standard error of 25 percent of
each source signature entry has been used strictly for illustrative pur-
poses. The printouts given in Tables C-6, C-7, and C-8 are the output
files from the interactive session.
167
-------
TABLE C-l. SOURCE NAME AND CODE INPUT FILE
1 SOIL
2 RD DUST
3 SEA SALT
4 SLSH BRN
5 AUTO CAT
6 AUTO
7 JET AIR
8 RES OIL
9 DIST OIL
10 RES 6AS
11 COAL
12 KRAFT RB
13 ELARCFRN
14 FERRMNFR
15 CAL GYP
16 CEMENT
17 PET FCC
18 LSKILN
19 SEC S04
168
-------
TABLE C-2. SPECIES NAME AND CODE INPUT FILE
1 TOT
2 AL
3 AS
4 BA
5 BR
6 CA
7 CD
8 CL
9 CO
10 CR
11 CU
12 FE
13 HG
14 K
15 MN
16 NI
1? P
18 PB
19 S04
20 SB
21 SI
22 SN
23 SR
24 TI
25 V
26 ZN
27 C
28 NA
29 N03
30 RB
31 SE
169
-------
TABLE C-3. FINE FRACTION SOURCE SIGNATURE INPUT FILE
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
••)
2
2
2
';>
2
•j
2
'•)
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
2V
18
.006500
.000000
.000000
.000000
.000000
.093000
.000000
.000000
.000000
.000300
.008300
.002000
.000370
.000000
.000000
.000000
.000140
.000020
.000000
.000000
.024000
.000000
.004500
.000280
.001400
.000000
.000220
.088400
.000000
.000000
.000200
.136500
.024400
.000000
.000000
.000450
.000300
.060000
.010300
.001230
.012500
.000090
.000000
.003700
.001625
.000000
.000000
.000000
.000000
.023250
.000000
.000000
.000000
.000075
.002075
.000500
.000093
.000000
.000000
.000000
.000035
.000005
.000000
.000000
.006000
.000000
.001125
.000070
.000350
.000000
.000055
.022100
.000000
.000000
.000050
.034125
.006100
.000000
.000000
.000112
.000075
.015000
.002575
.000307
.003125
.000022
.000000
.000925
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
m
NA
NI
NQ3
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
HN
NA
NI
H03
PB
FINE
FINE
FIHE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FIHE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
SOIL
SOIL
SOIL
SOU-
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
RH DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
K£i DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
170
-------
TABLE C-3 (CONTINUED)
2 30
2 20
2 31
2 21
2 22
2 19
2 23
2 24
2 25
2 26
3 2
3 3
3 4
3 5
3 27
3 6
3 7
3 8
3 10
3 11
3 12
3 14
3 15
3 28
3 16
3 29
3 18
3 30
3 20
3 31
3 21
3 22
3 19
3 23
3 24
3 25
3 26
4 2
4 3
4 4
4 5
4 27
4 6
4 7
4 8
4 10
4 11
.000000
.000000
.000000
.223000
.000000
.011100
.000000
.006400
.000230
.001100
.000000
.000000
.000000
.000000
.000000
.010000
.000000
.550000
.000000
.000000
.000000
.010000
.000000
.310000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.090000
.000000
.000000
.000000
.000000
.014400
.000000
.000530
.000530
.629000
.010700
.000530
.055500
.000000
.000900
.000000
.000000
.000000
.055750
.000000
.002775
.000000
.001600
.000057
.000275
.000000
.000000
.000000
.000000
.000000
.002500
.000000
.137500
.000000
.000000
.000000
.002500
.000000
.077500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.022500
.000000
.000000
.000000
.000000
.003600
.000000
.000132
.000132
.157250
.002675
.000132
.013875
.000000
.000225
RB
SB
SE
SI
SN
S04
SR
n
V
ZN
AL
AS
6A
BR
C
CA
CD
CL
CR
CU
FE
K
HN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
r'INE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
RD DUST
RD DUST
RD DUST
RIi DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
SEA SAL I
SEA SAL1
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SAL I
SEA SALT
SEA SALT
SEA SALT
SEA SAIT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
171
-------
TABLE C-3 (CONTINUED)
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
3
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
cr
,J
5
5
5
5
5
5
5
5
5
6
6
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
j
.001900
.006000
.001200
.006500
.000000
.051000
.000530
.000530
.000530
.000000
.008900
.000530
.016000
.000530
.000000
.000000
.000000
.001200
.000000
.000000
.000000
.390000
.001700
.000000
.000000
.000000
.000240
.001100
.000440
.000150
.000000
.000150
.000000
.000000
.000000
.000000
.000000
.005100
.000000
.500000
.000000
.000000
.000000
.000800
.000430
.000000
•
„
•
«
•
„
•
m
•
m
m
•
•
.
.
•
.
.
•
,
•
,
m
m
m
.
•
.
•
.
•
m
.
,
000475
001500
000300
001625
000000
012750
000132
000132
000132
000000
002225
000132
004000
000132
000000
000000
000000
000300
000000
000000
000000
097500
000425
000000
000000
000000
000060
000275
000110
000037
000000
000037
000000
000000
.000000
,
,
000000
000000
.001275
,
000000
.125000
.000000
.000000
.000000
.000200
.000108
.000000
FE
K
MN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
U
ZN
AL
ftS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
3N
504
SR
TI
V
U
AL
AS
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FIHE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUIO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
U CAT
U CAT
U CAT
y CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U LAf
U CAT
U CAT
U CA1
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CA!
U CAT
U CAT
172
-------
TABLE C-3 (CONTINUED)
6
6
6
A
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
f
7
/
;
/
7
/'
7
7
7
7
7
7
/
/
7
/
7
t
"7
/
7
7
7
-7
/
7
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
T?
C./
6
7
a
10
11
12
14
ID
28
16
2V
18
30
20
31
21
.000000
.082000
.545000
.000000
.000000
.054000
.000000
.000040
.002500
.000000
.000000
.000000
.000000
.000000
.211000
.000000
.000000
.000000
.000750
.000000
.002100
.000000
.000000
.000000
.000210
.000000
.000000
.000000
.000000
.960000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.020500
.136250
.000000
.000000
.013500
.000000
.000010
.000625
.000000
.000000
.000000
.000000
.000000
.052750
.000000
.000000
.000000
.000187
.000000
.000525
.000000
.000000
.000000
.000053
.000000
.000000
.000000
.000000
.240000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
BA
BR
c
CA
en
CL
CR
cu
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
MN
NA
HI
N03
PB
RB
SB
SI-
SI
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
AUTO
AUCO
AUTO
AUTO
AUTO
ftUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
JILT AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
173
-------
TABLE C-3 (CONTINUED)
7
7
7
7
7
7
8
8
8
8
8
8
8
a
8
8
8
8
a
8
8
a
8
a
8
8
a
8
8
8
8
8
8
9
9
9
9
9
9
9
9
9
9
9
9
9
9
22
19
23
24
25
26
2
3
4
5
27
6
7
a
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
.000000
.000000
.000000
.000000
.000000
.000000
.005300
.000000
.000000
.000130
.101000
.015800
.000000
.000000
.000470
.000750
.029700
.002800
.000460
.035000
.053600
.006500
.001100
.000000
.000000
.000000
.009600
.000000
.481000
.000000
.001100
.034400
.004000
.003100
.000000
.000000
.000260
.358000
.005000
.000000
.012000
.000000
.001700
.001200
.000180
.000140
.003200
.000000
.000000
.000000
.000000
.000000
.000000
.001325
.000000
.000000
.000033
.025250
.003950
.000000
.000000
.000117
.000187
.007425
.000700
.000115
.008750
.013400
.001625
.000275
.000000
.000000
.000000
.002400
.000000
.120250
.000000
.000275
.008600
.001000
.000775
.000000
.000000
.000065
.089500
.001250
.000000
.003000
.000000
.000425
.000300
.000045
.000035
.000800
SN
304
SR
TI
V
ZN
AL
AS
BA
SR
C
CA
CD
CL
CR
CD
FE
K
HN
NA
HI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
HN
NA
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JF.T AIR
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESII'UAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL.
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE UIL
174
-------
TABLE C-3 (CONTINUED)
V
9
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
11
11
11
11
11
n
11
16
29
18
30
20
31
21
22
19
23
24
25
26
'?
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
'T
3
4
5
27
A
7
.000090
.010000
.005400
.000000
,000000
.000000
.002700
.000000
.132000
.000000
.000000
.000050
.000290
.000000
.000000
.000000
.000000
.120000
.050000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.470000
.000000
.000000
.000000
.000000
.004380
.000000
.000120
.000000
.000000
.018180
,0008'->0
.000022
.002500
.001350
.000000
.000000
.000000
.000675
.000000
.033000
.000000
.000000
.000013
.000072
.000000
.000000
.000000
.000000
.030000
.012500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.117500
.000000
.000000
.000000
.000000
.001095
.000000
.000030
.000000
.000000
.004545
.000223
HI
N03
PB
REf
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
8R
C
CA
CD
CL
CR
CU
FE
K
hN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
Ab
BA
BR
C
CA
CD
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL.
DISTILLATE OIL.
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL 6AS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENHAL 8AS
RES I DEN! I AL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENHAL GAS
RES I DEN UAL GAS
COAL
COAL
COAL
COAL
COAL
COAL
COAL
175
-------
TABLE 03 (CONTINUED)
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
a
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
23
26
.000000
.000160
.000210
.043330
.001890
.000000
.002420
.000070
.000000
.006690
.000000
.000290
.000000
.035550
.008240
.096290
.000000
.001210
.000000
.016270
.002500
.000000
.000000
.001300
.019200
.000000
.000000
.018000
.002300
.000210
.012000
.015000
.000300
.127000
.001400
.000000
.OOOUO
.000000
.000000
.000000
.001500
.000000
.400000
.000000
.000060
.000010
.000690
.000000
.000040
.000053
.010832
.000472
.000000
.000605
.000018
.000000
.001673
.000000
.000072
.000000
.008888
.002060
.024072
.000000
.000303
.000000
.004068
.000625
.000000
.000000
.000325
.004800
.000000
.000000
.004500
.000700
.000053
.003000
.003750
.000075
.031750
.000350
.000000
.000033
.000000
.000000
.000000
.000375
.000000
.100000
.000000
.000015
.000002
.000173
CL
CR
CD
FE
K
MN
NA
HI
N03 "
PB
RB
SB
SE
SI
SN
504
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CO
CL
CR
CU
FE
K
MN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
FINE COAL
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
KRAFT REC
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER-
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
176
-------
TABLE C-3 (CONTINUED)
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
14
14
14
14
14
14
14
14
14
1 4
14
14
14
14
14
14
14
14
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
.006500
.000000
.000000
.000000
.000000
.062000
.000000
.018500
.021000
.002800
.319500
.009200
.087000
.012600
.007100
.000000
.007600
.000000
.000000
.000000
.050000
.000000
.025000
.000000
.002000
.000630
.012000
.006400
.000000
.000000
.001600
.105000
.013000
.000000
.004200
.000420
.000360
.021000
.105000
.173000
.031000
.000000
.057000
.000450
.000000
.001625
.000000
.000000
.000000
.000000
.015500
.000000
.004625
.005250
.000700
.079875
,002300
.021750
.003150
.001775
.000000
.001900
.000000
,000000
.000000
.012500
.000000
.006250
.000000
.000500
.000158
.003000
.001600
.000000
.000000
.000400
.026250
.003230
.000000
.001050
.000105
.000090
.005250
.026250
.043250
.007750
.000000
.014250
.000112
.000000
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
HN
NA
NI
N03
PB
RB
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINF;
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FIKE
FINE
FINE
ELECTRIC ARC FURN
ELECTRIC ARC FUR*
ELECTRIC ARC FURN
ELECTRIC ARC FURH
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
ELECTRIC ARC FURN
r-ERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGAiMESE FURN
FERROflANGANESE FURN
FERROMAN6ANESE FURN
FERROMARGViNESE FORK'
FERROrtANGANESE FURN
FERROrtANGANESE FURN
FERROMANGANESE I"-URN
FERROHANGANESE FURN
FERROMANGAIMESE FURN
FERROMANGANESE FURN
FERROMftNGANESE FURN
FERROMANGANESE I"URN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERRONANGANESE FURN
177
-------
TABLE C-3 (CONTINUED)
14
14
14
14
14
14
14
14
14
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
IS
15
15
16
la
16
16
16
16
16
16
16
16
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
J
4
5
27
6
7
8
10
11
.000000
.000000
.009900
.000000
.042000
.000000
.000460
.000240
.005800
.000000
.000000
.000000
.000000
.010000
.130000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.0.00000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.410000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.080000
.050000
.000000
.000000
.000000
.000000
.000000
.000000
.002475
.000000
.010500
.000000
.000115
.000060
.001450
.000000
.000000
.000000
.000000
.002500
.032500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.102500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.020000
.012500
.000000
.000000
.000000
.000000
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
c
CA
CD
CL
CR
cu
FE
K
MN
NA
HI
M03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
Cfl
CL
CR
CU
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROHANGANESE FURN
FERROMAN&ANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANIESE FURN
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAIQN OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GfPSUM
CALCINAION OF GYPSUM
CALCINAION OF BYPSUri
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION OF GYPSUM
CALCINAION Ol: GYPSUM
CALCINAION OF GYPSUM
CEMENT
CEMENT
CEMENT
CEMENT
CEHENF
CEMENT
CEHENI
CEMEN!
CEMENT
CEMENT
178
-------
TABLE C-3 (CONTINUED)
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
IB
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
9
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
.000000
.020000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.025000
.000000
.600000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.040000
.000000
.000000
.000000
.000000
.000000
.010000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.500000
.000000
.000000
.000000
.000000
.000000
.000000
.005000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.006250
.000000
.150000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.010000
.000000
.000000
.000000
.000000
.000000
.002500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.125000
.000000
.000000
.000000
.000000
.000000
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
SN
SQ4
SR
TI
V
ZN
AL
AS
BA
6R
c
CA
CD
CL
CR
cu
FE
K
MN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
II
V
ZN
AL
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE-
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
CEMENT
CEMENT
CEMENT
CEMENI
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
CEMENT
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PtTROL
PETROL
PETROL
PETROL
PETROL.
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PETROL
PE J-ROL
LS KILN
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
FCC
F CU
FCC
FCC
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITS
UNITb
UNITS
UNITS
179
-------
TABLE C-3 (CONTINUED)
18
18
18
18
18
18
18
18
18
1 8
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
,1
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
3
10
11
12
14
15
28
16
29
18
30
20
•
,
•
•
•
*
•
•
•
•
•
•
•
•
*
•
•
v
«
«
•
.
.
.
•
.
•
•
.
•
•
•
•
•
,
•
.
B
000000
000000
000000
400000
300000
000000
000000
000000
000000
020000
000000
000000
000000
000000
000000
000000
000000
000000
100000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
.
•
*
•
«
•
«
•
000000
000000
000000
100000
075000
000000
000000
000000
.000000
*
•
•
•
•
.
•
•
•
.
*
,
,
t
•
,
,
„
,
„
•
1
*
B
«
*
•
,
,
.000000
€
000000
.000000
,
„
000000
,000000
.000000
.000000
•
005000
000000
000000
000000
000000
000000
000000
000000
000000
025000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
000000
.000000
,
000000
.000000
.000000
.000000
AS
BA
BR
c
CA
CD
CL
CR
cu
FE
K
HN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
v
ZN
AL
AS
BA
BR
c
CA
CD
CL
CR
cu
FE
K
HN
NA
NI
N03
PB
RB
SB
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINE
FINk
FINE
FINE
FINE
FINE
KINE
FINE
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
L3 KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
SEC S04
SEC S04
SEC S04
SEC S04
SEC 304
SEC 304
SEC S04
SEC S04
SEC S04
SEC 304
SEC 304
SEC 304
SEC S04
SEC SO 4
SEC S04
SEC SO 4
SEC S04
SEC SO 4
SEC S04
180
-------
TABLE C-3 (CONTINUED)
19 31 .000000 .000000 SE FINE SE.C 304
19 21 .000000 .000000 SI FINE SEC S04
19 22 .000000 .000000 SN FINE SEC S04
19 19 1.000000 .000000 S04 FIXE SEC 304
19 23 .000000 .000000 SR FINE SEC 304
19 24 .000000 .000000 TI FINE SEC S04
19 25 .000000 .000000 V FINE SEC S04
19 26 .000000 .000000 ZN FINE SEC S04
181
-------
TABLE C-A. COARSE FRACTION SOURCE SIGNATURE INPUT FILE
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
'7
2
2
2
2
2
2
•7
2
2
9
2
y
2
-i
*.
••)
)
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
1 1
12
14
15
28
16
29
1(3
.005500
.000000
.000000
.000000
.000000
.051000
.000000
.000000
.000000
.000090
.004800
.001200
.000210
.000000
.000000
.000000
.000200
.000020
.000000
.000000
.024000
.000000
.002700
.000140
.000890
.000000
.000090
.065900
.000000
.000000
.000080
.048900
.030000
.000000
.000000
.000000
.000000
.057300
.000000
.001000
.017500
.000000
.000200
.000000
.001375
.000000
.000000
.000000
.000000
.012750
.000000
.000000
.000000
.000022
.001200
.000300
.000053
.000000
.000000
.000000
.000050
.000005
.000000
.000000
.006000
.000000
.000675
.000035
.000223
.000000
.000022
.016475
.000000
.000000
.000020
.012225
.007500
.000000
.000000
.000000
.000000
.014325
.000000
.000250
.004375
.000000
.000050
.000000
AL
AS
BA
BR
C
CA
CD
CL
CR
CD
FE
K
MN
HA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CD
FE
K
HN
NA
NI
H03
PB
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
l.OARSE
SOIL
SOU-
SOIL
SOU-
SOIL
SOU-
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOIL
SOU-
SOIL
SOIL
SOIL
SOU-
SOIL
RD DUST
RD DUST
RD OUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD OUST
182
-------
TABLE C-4 (CONTINUED)
2 30
2 20
2 31
2 21
2 22
2 19
2 23
2 24
2 25
2 26
3 2
3 3
3 4
3 5
3 2?
3 6
3 7
3 8
3 10
3 11
3 12
3 14
3 15
3 28
3 16
3 29
3 18
3 30
3 20
3 31
3 21
3 22
3 19
3 23
3 24
3 25
3 26
4 2
4 3
4 4
4 5
4 27
4 6
4 7
4 8
.000000
.000000
.000000
.284000
.000000
.000000
.000000
.010100
.000270
.000000
.000000
.000000
.000000
.000000
.000000
.010000
.000000
.550000
.000000
.000000
.000000
.010000
.000000
.310000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.090000
.000000
.000000
.000000
.000000
.002300
.000000
.000000
.000000
.523000
.001300
.000000
.009300
.000000
.000000
.000000
.071000
.000000
.000000
.000000
.002525
.000068
.000000
.000000
.000000
.000000
.000000
.000000
.002500
.000000
.137500
.000000
.000000
.000000
.002500
.000000
.077500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.022500
.000000
.000000
.000000
.000000
.000575
.000000
.000000
.000000
.130750
.000325
.000000
.002325
RB
SB
SE
SI
SN
S04
SR
TI
U
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CD
FE
K
HN
NA
NI
NQ3
PB
R6
SB
SE
SI
SIN!
S04
SR
n
V
ZN
AL
AS
BA
BR
C
CA
CU
CL
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COftRSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
RD DUS!
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RD DUST
RB DUST
RD DUST
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SALT
SEA SAL!
SEA SALT
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURiV
SLASH BURN
SLASH BURN
SLASH BURN
SLASH BURN
183
-------
TABLE C-4 (CONTINUED)
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
&
5
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
.000000
.000000
.000000
.000000
.000090
.001600
.000000
.010000
.000000
.000000
.000000
.000000
.054000
.000000
.007500
.000000
.000000
.000000
.000000
.001200
.000000
.000000
.000000
.390000
.001700
.000000
.000000
.000000
.000240
.001100
.000440
.000150
.000000
.000150
.000000
.000000
.000000
.000000
.000000
.005100
.000000
.500000
.000000
.000000
.000000
.000800
.000000
.000000
.000000
.000000
.000022
.000400
.000000
.002500
.000000
.000000
.000000
.000000
.013500
.000000
.001875
.000000
.000000
.000000
.000000
.000300
.000000
.000000
.000000
.097500
.000425
.000000
.000000
.000000
.000060
.000275
.000110
.00003?
.000000
.000037
.000000
.000000
.000000
.000000
.000000
.001275
.000000
.125000
.000000
.000000
.000000
.000200
CR
CD
FE
K
NN
HA
NI
N03
PB
Rfi
SB
SE
SI
SN
804
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CD
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
CUARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
SLASH
AUTO
AUTO
AUTO
AUTO
AUFO
AUFO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUIO
AUTO
AUTO
AUTO
AUTO
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
BURN
U CA!
U CAT
U CAT
U CAI
U CAI
U CAT
U CAI
U CAI
U CAT
U CAI
U CAI
U CAT
U CAT
U CAI
U CAT
U CAI
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAT
U CAI
U CAI
U CAT
U CA1
184
-------
TABLE C-4 (CONTINUED)
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
/'
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
j
'•)
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
9
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
.000430
.000000
.000000
.082000
.545000
.000000
.000000
.054000
.000000
.000040
.002500
.000000
.000000
.000000
.000000
.000000
.211000
.000000
.000000
.000000
.000750
.000000
.002100
.000000
.000000
.000000
.000210
.000000
.000000
.000000
.000000
.960000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000108
.000000
.000000
.020500
.136250
.000000
.000000
.013500
.000000
.000010
.000625
.000000
.000000
.000000
.000000
.000000
.052750
.000000
.000000
.000000
.000187
.000000
.000525
.000000
.000000
.000000
.000053
.000000
.000000
.000000
.000000
.240000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
AL
AS
BA
BR
c
CA
CD
CL
CR
cu
FE
K
«N
HA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
UN
HA
HI
N03
PB
RB
SB
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JE1 AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JEI AIR
JEI AIR
185
-------
TABLE C-4 (CONTINUED)
7 31
7 21
7 22
7 19
7 23
7 24
7 25
7 26
8 2
8 3
8 4
8 5
8 27
8 6
8 7
8 8
8 10
8 11
8 12
8 14
8 15
8 28
8 16
8 29
8 18
8 30
8 20
8 31
8 21
8 22
8 19
8 23
8 24
8 25
8 26
9 2
9 3
9 4
9 5
9 27
9 6
9 7
9 8
9 10
V 11
9 12
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.014700
.000000
.000000
.000000
.315000
.035600
.000000
.007400
.000540
.000860
.014700
.000000
.000270
.009000
.013900
.000000
.000000
.000000
.000000
.000000
.037000
.000000
.073000
.000000
.001400
.017900
.000500
.005900
.000000
.000000
.000000
.172000
.005000
.000000
.016000
.000000
.001100
.002300
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.003675
.000000
.000000
.000000
.078750
.008900
.000000
.001850
.000135
.000215
.003675
.000000
.000068
.002250
.003475
.000000
.000000
.000000
.000000
.000000
.009250
.000000
.018250
.000000
.000350
.004475
.000125
.001475
.000000
.000000
.000000
.043000
.001250
.000000
.004000
.000000
.000275
.000575
SE
SI
SN
SQ4
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
MN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
JET AIR
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL.
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
RESIDUAL OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE UIL
DISTILLATE UIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL.
DIS1ILLATE OIL
DI31ILLATE OIL
DIS1ILLATE OIL
DISTILLAIE UIL.
186
-------
TABLE C-4 (CONTINUED)
9
y
9
9
9
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
14
15
28
16
2?
18
30
20
31
21
22
19
23
24
25
26
<-)
3
4
5
27
6
7
a
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
2S
24
25
26
.000000
.000130
.006900
.000000
.000000
.000000
.000000
.000000
.000000
.002900
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.120000
.050000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.470000
.000000
.000000
.000000
.000000
.000000
.000033
.001725
.000000
.000000
,000000
.000000
.000000
.000000
.000725
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.030000
.012500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.117500
.000000
.000000
.000000
.000000
K
M
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
c
CA
CD
CL
CR
CU
FE
K
HN
NA
NI
N03
PB
RB
SB
SE
SI
SN
304
5R
TI
v
ZN
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
1HSTILLAIE OIL
DISTILLATE OIL.
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILL ATE "OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
DISTILLATE OIL
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDEN1IAL GAS
RESIDENTIAL SAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL 6AS
RESIDENTIAL GAS
RESIDE.NIIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RESIDENTIAL GAS
RES10ENUAL GAS
RESIDtNIIAL GAb
187
-------
TABLE C-4 (CONTINUED)
11
1 1
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
?
8
10
11
12
14
15
28
16
2V
18
.000000
.000000
.000000
.000000
.000000
.023600
.000000
.000000
.000000
.000000
.040990
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.032920
.000000
.121120
.000000
.000000
.000000
.006830
.002800
.000000
.000000
.000560
. 1 76000
.003600
.000000
.029000
.004800
.000600
.018400
.004000
.000520
.053000
.002200
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.005900
.000000
.000000
.000000
.000000
.010248
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.008230
.000000
.030280
.000000
.000000
.000000
.001708
.000700
.000000
.000000
.000140
.044000
.000900
.000000
.007250
.001200
.000150
.004600
.001000
.000130
.013250
.000550
.000000
.000000
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
m
NA
NI
«03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
UN
HA
NI
N03
'rii
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COAKSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COftRSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
COAL
KRAF T
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAF I
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
KRAFT
REC
REC
REC
REC
REC
REC
REC
REC
KEC
REC
REC
REC
RiiC
REC
REC
REC
REC
BUILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
flOILER
BOILER
BUILER
BOILER
BOILER
BOILER
BOILER
BOILER
BOILER
188
-------
TABLE C-4 (CONTINUED)
\2
12
12
12
12
12
12
12
12
12
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
14
14
14
14
14
14
14
14
14
14
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
3
27
6
7
8
10
11
.000000
.000000
.000000
.001300
.000000
.118000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.069000
.000000
.038000
.021000
.004300
.308000
.005600
.095000
.014000
.006200
.000000
.006400
.000000
.000000
.000000
.066000
.000000
.013000
.000000
.001100
.000580
.012000
.001800
.000000
,000000
.000000
.190000
.005,700
.000000
.008500
.000230
.000740
.000000
.000000
.000000
.000325
.000000
.029500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.017250
.000000
.009500
.005250
.001075
.077000
.001400
.023750
.003500
.001550
.000000
.001600
.000000
.000000
.000000
.016500
.000000
.003250
.000000
.000275
.000145
.003000
.000450
.000000
.000000^
.000000"
.047500
.001425
.000000
.002125
.000057
.000185
KB
SB
SE
SI
SN
S04
SR
ri
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
«N
NA
MI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
cn
CL
CR
CU
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
KRAFT REC BOILER
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC' FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC I-URN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
ELEC ARC FURN
FERROMANGANESE. FURN
FERROMANGANF.5E FURN
FERROMANGANESE' FURN
FERROrtANGANESE FURN
FERRQMANGANIISE FURN
FERROiiArtGANESE FURN
FERROMANGANESE FURN
FERROMANGANhSE. FURN
FERROMANGANESE FURN
FERROMANGANESE FUIVN
189
-------
TABLE C-4 (CONTINUED)
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
1 5
15
15
15
15
16
16
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
•>
3
.003800
.012000
.027500
.003400
.000000
.012000
.000000
.000000
.000000
.000000
.002900
.000000
.011000
.000000
.000000
.000090
.000350
.000000
.000000
.000000
.000000
.000000
.090000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.5/0000
.000000
.000000
.000000
.000000
.000000
.000000
.000950
.003000
.006875
.000850
.000000
.003000
.000000
.000000
.000000
.000000
.000725
.000000
.002750
.000000
.000000
.000022
.000087
.000000
.000000
.000000
.000000
.000000
.022500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.142500
.000000
.000000
.000000
.000000
.000000
.000000
.FE
K
MN
NA
NI
N03
PB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
HN
NA
NI
H03
PB
RB
SB
SE
SI
SN"
S04
SR
TI
V
ZN
AL
AS
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
FERROMANGANESE FURN
FERROMANGANE.SE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERRQMANGANESE FURN
FERROMANGANESE FURN
FERRQMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERRQMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
FERROMANGANESE FURN
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINAIION OI-" GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CALCINAIION OF GYPSUM
CALCINATION OF GYPSUM
CALCINATION OF GYPSUM
CEMEN1
CEMENT
190
-------
TABLE C-4 (CONTINUED)
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
17
17
17
17
17
17
17
17
17
1 7
17
17
1 7
17
17
17
17
17
17
17
17
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
2
3
4
'5
27
6
7
8
10
11
12
14
15
-------
TABLE 04 (CONTINUED)
17
17
1,'
17
17
17
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
18
19
19
19
19
19
19
19
19
19
19
19
19
19
19
22
19
23
24
25
26
2
3
4
5
27
6
7
8
10
11
12
14
15
28
16
29
18
30
20
31
21
22
19
23
24
25
26
o
3
4
5
27
6
7
8
10
11
12
14
15
26
.000000
.070000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.400000
.300000
.000000
.000000
.000000
.000000
.020000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.100000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.017500
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.140000
.075000
.000000
.000000
.000000
.000000
.005000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.025000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
SN
S04
SR
TI
V
ZN
AL-
AS
BA
BR
c
CA
CD
CL
CR
cu
FE
K
UN
NA
NI
N03
PB
RB
SB
SE
SI
SN
504
SR
TI
V
ZN
AL
AS
BA
BR
C
CA
CD
CL
CR
CU
FE
K
MN
Nfi
PETROL-
PETROL-
PETROL-
PETROL-
PETROL-
PETROL -
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
FCC UNDTb
•FCC UNO IS
FCC UMTS
•FCC UNDTS
•FCC UNDTS
FCC UNDTS
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILH
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
LS KILN
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY 304
SECONDARY S04
SECONDARY SU4
SECONDARY S04
SECONDARY SU4
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY SIM
SECONDARY S04
192
-------
TABLE C-4 (CONTINUED)
19
19
19
19
19
19
19
19
19
19
19
19
19
16
29
18
30
20
31
21
22
19
23
24
25
26
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
1,000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
.000000
HI
N03
FB
RB
SB
SE
SI
SN
S04
SR
TI
V
ZN
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
COARSE
SECONDARY S04
SECONDARY 304
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY SO4
SECONDARY S04
SECONDARY S04
SECONDARY S04
SECONDARY 804
SECONDARY 304
SECONDARY 304
193
-------
TABLE C-5. RECEPTOR CONCENTRATION INPUT FILE
03 URBftN CORE 810229 12 07
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
03 BACKGROUND 81022? 12 07
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
35.
.088
.036
.149
.344
.441
.004
.051
.003
.002
.013
.208
0.
.165
.022
.006
.119
1.422
13.13
.003
.277
.001
.003
.007
.002
.042
28.
.070
.014
.146
.164
.128
.001
.010
.008
.010
.027
.120
.0015
.186
.022
.003
7.
.062
.007
.014
.017
.023
.002
.007
.002
.004
.002
.011
.003
.009
.004
.001
.017
.068
.681
.002
.031
.002
.001
.014
.008
.003
3.
.055
.006
.018
.010
.008
.002
.007
.003
.005
.003
.008
.003
.010
.004
.002
75.
1.951
.004
.371
.108
13.014
.0005
.288
.015
.022
.018
1.871
.001
.517
.088
.003
.102
.346
1.725
.004
8.652
.005
.030
.205
.007
.071
37.
.739
.001
.086
.064
3.805
.001
.014
.001
.0025
.004
.692
.009
.215
.018
.001
7.
.422
.004
.023
.007
1.442
.001
.068
.003
.005
.002
.118
.002
.072
.008
.001
.040
.024
.939
.002
1.987
.002
.003
.037
.014
.006
4.
.171
.004
.016
.005
.444
.002
.010
.003
.005
.002
.044
.003
.032
.004
.002
194
-------
TABLE C-5 (CONTINUED)
30 17 .103 .018 .020 .022
30 18 .894 .047 .213 .017
30 19 14.78 .765 .852 .783
30 20 .008 .003 .001 .003
30 21 .189 .027 4.0 .914
30 22 .024 .003 .003 .003
30 23 .003 .002 .006 .002
30 . 24 .003 .016 .097 .017
30 25 .003 .009 .0035 .007
30 26 .038 .003 .016 .003
195
-------
w
hJ
M
EH
H
O<
H
O
SH
O,
O
U
s
p
C/3
w
I
en
i
en CD ;c
UJ » =J
cc •— o
— — «r CN r\. •— o
fx tf? rj ^ ^r
>4 I 1 I I I I I
CM CN »- ^- ro r*i in «-
— rs.— rx
oo-ior-jrxi-oiooi
CO GO O-
O rs. — ^- XJ CM
§O"Oinoor4OO
O--^OOr*>OO
00*0 ro
O^TtO — rofv,OOm — •^•ir3f*5O-ooooa---ot>j
^x O
-o o
CN O
0
o
> o
196
-------
Q
W
2!
o
u
«r u*j Pi CO r*3 —* --a rj •— 00 c%i n rv u"i
c»o — m c*j K) mrv.ro
rt r~i w—
C ^03i*"JOrv-orN.r*.rjij"3 *•• •— o* li") *r 'O ~-
<&oo*— OOoortOO'OO1
5^iO
*— ftg u^
O
O -C I — — S3
LJ •• oo
u. *— uj
O- v— OO Irt ^ *O O
CM (NI
en vi
tn tn
C_3 C/) •« £_» O «T Ci.
U)
197
-------
TABLE C-7. GENERAL HAKDCOPY OUTPUT FILE
RESULTS FOR CUB SITE: URBAN CORE YEAR: 81 DATE: 0229
COARSE PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. U1TH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RIDGE REGRESSION LEAST SQUARES
K= .800
CODE
SOURCE
INTERCEPT
1
2
3
5
6
9
10
11
12
13
14
16
17
19
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
DIST OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRMNFR
CEMENT
PET FCC
SEC S04
TOTAL:
SPECIES FIT
CODE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FLG
TOTAL *
AL *
AS
BA «
BR *
CA *
CD
CL *
CO *
CR *
CU »
FE *
HG
K *
MN *
NI
P *
PB *
S04
St
SI *
SN
SR *
TI *
V
ZN *
C
NA
N03
RB
SE
UG/M3
.011+-
34.445+- 18
13.578+- 4
.177+-
10.339+- 13
.722+-
.000+-
.000+-
4.532+- 3
2.229+- 3
.000+-
.000+-
13.874+- 11
.000+-
.000+-
81.908+- 25
.005
.752
.185
.491
.655
.275
.000
.000
.137
.455
.000
.000
.324
.000
.000
.267
KISS COftRSE
FLG HEAS. UG/M3
75.000+- 7
1.951+-
.004+-
.371+-
.108+-
13.014+- 1
C
.288+-
.015+-
.022+-
.018+-
1.871+-
<
.517+-
.088+-
.003+-
.102+-
.344+-
1.725+-
.004+-
8.452+- 1
.005+-
.030+-
.205+-
.071*-
M
N <
M
M <
M
.000
.422
.004
.023
.007
.442
.001
.068
.003
.005
.002
.118
.001
.072
.008
.001
.040
.024
.939
.002
.987
.002
.003
.037
.007
.004
.000
.000
.000
.000
.000
R-SQUARE:
Z
48.593*-25
18.1Q4+- 8
.328*-
9.124*-18
.963*-
.000*-
.000+-
6.043+- 4
2.308+- 4
.000+-
.000+-
18.499+-I5
.000+-
.000+-
109.210+-35
.6151
K-0.0
R-SQUARE:
UG/M3
.410
.418
.655
.252
.378
.000
.000
.220
.615
.000
.000
. 1 97
.000
.000
.197
77
18
-51
1
11
-
23
81
.023+-
.894+-
.281+-
.365+-
.096*-
.180+-
.000+-
.000+-
.906+-
.031+-
.000+-
.000+-
.364+-
.000+-
.000+-
.884*-
.010
81.347
13.310
.771
63.596
.395
.000
.000
7.909
6.302
.000
.000
20.317
.000
.000
43.046
6
2
1.
6
1
2
1
1
.7702
VIF
.178
.184
.223
.754
.054
.000
.000
.32?
.564
.000
.000
.550
.000
.000
SUSPENDF.ll PARTICIPATE
PERCENT
100.000+-13.
2.601+- .
.003+- .
.495+- .
.144+- .
17.352+- 2.
<
.384+- .
.020+- .
.029+- .
.024+- .
2.495+- .
<
.689+- .
.117+- .
.004+- .
.134*- .
.461*- .
2.300+- 1.
.005+- .
11.336+- 2.
.007+- .
.040+- .
.273+- .
19?
613
005
055
016
514
001
097
004
007
003
281
003
116
015
001
055
054
270
003
860
003
005
056
< .019
.095+- .012
< .001
< .001
< .001
f .001
< .001
CAI.C. UG/M3
81.908+-25
1.125+-
.011+-
.011+-
.073+-
5.I86+-
.011+-
.212+-
.011+-
.022+-
.018+-
1.204+-
.011*-
.347+-
.035*-
.018+-
.011 *-
.171+-
6.525+- 1
.om-
4.335+- 1
.011+-
.016+-
.181+-
.015+-
.054+-
8.129+- 1
.422+-
.014+-
.012*-
.011+-
.267 1
.229
.000 2
,000
.015
.842
.000 22
.031
.000
.003
,001 1
.205
.000 11
.070
.004
.001 5
,000
.038
.306 3
.000 2
.048
.000 2
.001
.035
.001 2
.008
.224
.068
.001
.000
.000
RATIO
.092+-
.577+-
.762+-
.030+-
.672+-
.399+-
.097+-
.73i+-
.737*-
.9S9+-
.010+-
.643+-
.048+- 4
.672+-
.398+-
.834+- 2
.108+-
.494+-
.782+- 2
.762+-
.732+-
.210+-
.53B+-
.881+-
.102+-
.756+-
.000+-
.000*-
.000+-
.000+-
.000+-
.499
.136
.000
.000
.165
.070
.000
.133
.035
.171
.084
.130
.756
.164
.048
.534
.004
.123
.962
.000
.150
.000
.048
.229
.307
.142
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CU
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
S04
SB
SI
SN
SR
TI
V
IN
C
NA
N03
RB
SE
HEASURED AMBIENT MASS (UG/N3): FINE: 35.0+- 7.0 COARSE: 75.0+- 7.0 TOTAL: 110.0*- 9.9
L98
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CMS SHE: URIAH CORE TEAR: 81 DATE: 0229
FINE PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. UITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RIDGE REGRESSION
K= .010 R-SQUARE: .9179
CODt
SOURCE
INTERCEPT
1
2
3
5
6
9
10
1 1
13
14
16
17
IV
SOU
RD DUST
SEA SALT
AUTO CAT
AUTO
DIST OIL
RES CAS
COAL
ELARCFRN
FERRHNFR
CEMENT
PET FCC
SEC S04
TOTAL:
SPECIES FIT
CODE
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FLG
TOTAL *
AL *
AS *
DA *
BR *
CA *
CD *
CL »
CO »
CR
CU *
FE *
HQ
K »
HN »
NI *
P »
PB *
S04 «
SB *
SI »
SN
SR «
II
V
ZN *
C
NA
N03
RB
SE
UG/M3
.004+-
-.50?+-
-.024+-
-.446+-
7.377+-
4.635+-
3. 596+-
1 .0)5+-
1 .77H-
.076+-
.041+-
7.403+-
8.22I+-
-.187+-
32.991+-
KISS
FLG HEAS.
35.000+-
.088+-
.034+-
.149+-
.344+-
.441+-
.004+-
.051+-
.003+-
.013+-
.208+-
<
.165+-
.022+-
.006*-
.119+-
I.422+-
13.130+-
.003*-
.277*-
<
.003*-
.042+-
(1
h <
H
H <
(1
.004
3.092
.923
.171
20.436
.941
6.6.14
5.311
1.348
.454
.250
2.919
12.252
11.012
15.162
FINE
UG/M3
7.000
.062
.00;
.014
.017
.023
.002
.007
.002
.002
.002
.011
.000
.00?
.004
.001
.01/1
.068
.681
.002
.031
.001
.001
.007
.002
.003
.000
.000
.000
.000
.000
X
4.4m- 8.838
1.318+- 2.636
.276+- .552
29. 271 +-58.541
13.244*- 3.773
9.305+-19.009
7.593+-15.186
5.059+- 3.981
.648+- 1.296
.358+- .715
21.151+- 9.353
17.660+-35.320
15.732+-31.444
94.261+-47.244
LEAST
K-0.0
SQUARES
R-SOUARE:
UG/M3
-8
-
-2
4
6
15
2
7
9
-3
32
.005+-
.858+-
.891+-
.533+-
.525+-
.800+-
.549+-
.427+-
.192*-
.048+-
.083+-
.999+-
.410+-
.353-+-
.136+-
.006
27.313
3.239
.265
58.257
.975
10.723
49.140
2.462
1 .249
.612
4.322
41 .216
49.346
50.953
432
29
14
79
i
14
413
9
71
24
6
77
173
.9200
VIF
.869
.638
.234
.398
.881
.366
.108
.374
.640
.698
.798
.121
.613
SUSPENDED PARTICULATE
PERCENT
100.000+-28.284
.251*- .184
.103+- .029
.424+- .094
.983+- .202
1.260+- .260
.011+- .006
.144+- .035
.009+- .006
; .01!
.037+- .009
.5941- .123
< .009
.471+- .098
.063+- .017
.017+- .004
.340+- .084
4.063+- .835
37.514+- 7.751
.009+- .006
./91+- .181
< .006
.009+- .003
< .040
( .023
.120*- .025
< .003
< .003
C .003
< .003
< .003
CALC.
UG/H3
32.991+-15
.0291-
.004+-
.004+-
.385+-
.441+-
.006+-
.054+-
.004+-
.006+-
.012+-
.207+-
.0041-
.161+-
.022+-
.006+-
.004+-
1 .014+-
13.151+-
.004+-
.290+-
.019+-
.004+-
.005+-
.004+-
.042+-
7.740+-
-.116+-
.043+-
.004+-
.004+-
1
1
.162
.004
.000
.000
.095 1
.095
.000 1
.088 1
.000 1
.000 2
.002
.029
.000
.037
.003 1
.000 1
.000
.245
.780 1
.000 1
.050 1
.004 18
.000 1
.001
.000 2
.007
.024
.035
.009
.000
.000
RATIO
.943+-
.331+-
.109+-
.028+-
.119+-
.999+-
.375+-
.049+- 2
.308+-
.912+-
.957+-
.994+-
.000+-
.?73+-
.019+-
.003+-
.033+-
.713+-
.002*-
.479*-
.045+-
.5I6+-67
.261+-
./6a*-
.080+-
.99^+-
.000+-
.000+-
.000+-
.000+-
.000+-
.595
.050
.000
.000
.415
.303
.168
.510
.081
.626
.170
.197
.000
.314
.202
.075
.001
.211
. 192
.076
.761
.641
.019
.10J
.056
.247
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
5U4
SB
SI
5N
SR
n
V
ZN
C
NA
N03
RB
SE
MEASURED AHUHNl HASS (UG/M3): HNE: 35.0 + - 7.0 COARSE: ,'5.0 + - 7.0 TOTAL: 110.0*- 9.9
199
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CM SITE: URBAN CORE TEAR: 81
COARSE PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. UITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RIDGE REGRESSION
K- .800 R-SQUARE:
CODE
SOURCE
INTERCEPT
1
2
3
5
4
?
10
11
12
13
14
14
17
1?
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
DIST OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRHNFR
CEMENT
PET FCC
SEC S04
TOTAL:
SPECIES FIT
CODE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FLG
TOTAL *
AL *
AS
BA *
BR *
CA *
CD
CL *
CO *
CR *
CU *
FE *
HG
K «
NH *
HI
P «
PB *
S04
SB
SI *
SN
SR »
TI »
V
ZN »
C
NA
N03
RB
SE
UG/H3
.011+-
36.445+- 18
13.578*- 6
.177+-
10.339+- 13
.722+-
.000+-
.000+-
4.532+- 3
2.229+- 3
.000+-
.000+-
13.874+- 11
.000+-
.000+-
81.908+- 25
.005
.752
.185
.491
.655
.275
.000
.000
.137
.435
.000
.000
.324
.000
.000
.247
HISS COARSE
FLG KEAS. U6/M3
75.000+- 7
1.951+-
.004*-
.371+-
.108+-
13.014+- 1
^
.288+-
.015+-
.022+-
.018+-
1 .871+-
<
,517+-
.088+-
.003+-
.102*-
.346+-
1.725+-
.004+-
8.452+- t
.005+-
.030+-
.205+-
(
.071+-
H <
H <
M i
fl <
M <
.000
.422
.004
.023
.007
.442
.001
.068
.003
.005
.002
.118
.001
.072
.008
.001
.040
.024
.939
.002
.987
.002
.003
.037
.007
.006
.000
.000
.000
.000
.000
I
48.593+-2S
18.104+- 8
.328*-
9.126+-18
.963+-
.000*-
.000+-
6.043+- 4
2.308+- 4
.000+-
.000+-
I8.499+-15
.000*-
.000+-
109.210+-35
DATE: 0229
.6151
LEAST SQUARES
K-0.0 R-SQUARE:
UG/M3
.410
.418
.655
.252
.378
.000
.000
.220
.615
.000
.000
.197
.000
.000
.197
77
18
-51
1
11
-
23
81
.023+-
.894+-
.281+-
.365*-
.096+-
.180+-
.000+-
.000*-
.906*-
.031*-
.000*-
.000+-
.364+-
.000-+-
.000+-
.884*-
.010
81.347
13.310
.771
63.576
.395
.000
.000
7.909
6.302
.000
.000
20.317
.000
.000
43.046
6
2
1
6
1
2
1
1
.7702
VIF
.173
.184
.223
.754
.054
.000
.000
.329
.564
.000
.000
.550
.000
.000
SUSPENDED PARTICULATE
PERCENT CALC. UG/H3
100.000+-13.
2.601+- .
.003+- .
.495+- .
.144*- .
17.352*- 2.
<
.384+- .
.020+- .
.029+- .
.024+- .
2.495+- .
<
.689+- .
.117+- .
.004*- .
.136+- .
.461+- .
2.300+- 1.
.005+- .
11.536+- 2.
.007+- .
.040+- .
.273+- .
<
.095+- .
<
<
f
^ .
.s
\ .
(
199 81.
613 1.
005
055
016
514 5.
001
097
004
007
003
281 1.
003
116
015
001
055
054
270 6.
003
860 4.
003
005
056
019
012
001 8.
001
001
001
001
908+-25
I25+-
011+-
011 *-
073+-
186+-
011+-
212*-
011*-
022+-
018+-
204+-
011 +-
347*-
035-*-
018+-
01I+-
171+-
525*- 1
011+-
335*- 1
011+-
016+-
I8I+-
015+-
054*-
12V+- 1
422+-
014+-
012+-
011+-
.267 1
.229
.000 2
.000
.015
.842
.000 22
.031
.000
.003
.001 1
.205
.000 11
.070
.004
.001 5
.000
.038
.306 3
.000 2
.048
.000 2
.001
.035
.001 2
.008
.224
.068
.001
.000
.000
RATIO
.092f-
.377*-
.762*-
.030+-
.672+-
.399+-
.097+-
.736+-
.737+-
.989+-
.0\0+-
.643+-
.048+- 4
.672+-
.398+-
.834+- 2
.108+-
.494+-
.782+- 2
.762+-
.732+-
.210+-
.538+-
.881+-
.102+-
.-'56+-
.000+-
.000+-
.000+-
.000+-
.000+-
.499
.134
.000
.000
.165
.070
.000
.133
.035
.171
.084
.130
.756
.164
.048
.534
.004
.123
.962
.000
.130
.000
.048
.229
.307
.142
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
MASS (UG'MI): FINE: 35.0 + - 7.0 COAP?F: 7-5.0*- 7.0 IPlAl: 110.0'
200
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CNB SUE: URJAK CORE
TOTAL PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. UITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RESULTS DERIVED FROM FINE AND COARSE FITTINGS
fEAR: 8! DATE: 022?
CODE
1
2
3
5
6
9
10
11
12
13
14
16
17
1?
SOURCE UG/W3
INTERCEPT .015+- .006
SOIL 35.936+- 19.005
RD DUST 13.554+- 6.254
SEA SALT -.270+- .520
AUTO CAT 17.717+- 24.579
AUTO 5.358+- .980
DIST OIL 3. 594+- 6.614
RES GAS 1.015*- 5.311
COAL 6.303+- 3.414
KRAFT RB 2.229+- 3,455
ELARCFRN .076+- .454
FERRHNFR .061+- .250
CEMEHT 21.277+- 11.694
PET FCC 8.221+- 12.252
SEC S04 -.187+- 11.012
TOTAL: 114. 897+- 27.467
SPECIES FIT HISS TOTAL
CODE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
14
17
18
1?
20
21
22
23
24
25
26
27
28
29
30
31
FL6 FL6 HEAS. UG/M3
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
HN
HI
P
ft
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
* 110,000+-
* 2.039+-
.040+-
* ,520+-
* .452+-
* 13.455+-
.004+-
* .339+-
* .018+-
* ' .024+-
* .031+-
* 2.079+-
<
» .682+-
* .110+-
.00?+-
* .221+-
* 1.748+-
14.855+-
.007*-
* 8.92?*-
.006+-
» .033+-
* .212+-
<
* .113+-
N <
H <
M
H <
M <
9.899
.427
.008
.027
.018
1 .442
.002
.066
.004
.006
.003
.11?
.001
.073
.00?
.001
.043
.072
1.160
.003
1.?87
.003
.003
.040
.00?
.007
.000
.000
.000
.000
.000
I
32.669+-17.525
12.321+- 5.792
.237+- .473
11.196+-22.391
4.871+- .973
3.010+- 6.020
2.415+- 4.829
3.730+- 3.146
1.573+- 3.146
.206+- .412
.114+- .227
19.343+-10.773
5.579+-11.159
5.006+-10.011
104.454+-28.390
SUSPENDED PARTICULAR
PERCENT
100.000+-12
1.854-+-
.036+-
.*73+-
.411+-
12.232-1- 1
.004+-
.308+-
.014+-
.022+-
.028+-
1.870+-
(
.620+-
.100+-
.008+-
.201+-
!. 407-1-
13.505+- 1
.004*-
8.1 17+- 1
.005+-
.030+-
.193*-
<
.103+-
<
<
(
<
<
.727
.422
.008
.049
.041
.712
.002
.068
.004
.006
.004
.201
.003
.086
.012
.001
.043
.139
.60?
.003
.?4?
.003
.004
.040
.015
.011
.001
.001
.001
.001
.001
CALC. UG/M3
114.899*-29
1.154+-
.015+-
.015+-
.458*-
5.627+-
.017+-
.266+-
.015+-
.028+-
.031+-
1.411+-
.015+-
.508+-
.057-+-
.024+-
.015+-
l.!84+-
1?.676+- 2
.015+-
4.624+- 1
.030+-
.020+-
.186+-
.019+-
.0?5+-
15.870+- 1
.305+-
.057+-
.016+-
.015+-
.467
.22?
.000
.000
.096
.847
.000
.094
.000
.003
.002
.207
.000
.079
.005
.001
.000
.248
.208
.000
.050
.004
.001
.035
.001
.011
.596
.076
.009
.000
.000
RATIO
1 .045+-
.544+-
.374+-
.029+-
1.013+-
.418+-
3.677+-
.78J+-
.832+-
1.I49+-
.988+-
.679+-
14.973+-
.745+-
.522+-
2.613+-
.068+-
.670+-
1.325+-
2.212+-
.742+-
4.927+-
.604+-
.877+-
2.097+-
.845+-
.000+-
.000+-
.000+-
.000+-
.000+-
.387
.129
.000
.000
.303
.068
.334
-.350
.033
.172
.087
.120
6.801
.145
.051
.411
.002
.149
.247
.044
.146
3.057
.045
.221
.239
.124
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
m
NI
P
FB
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
MEASURED AMBIENT HASS (UG/M3): FINEi 35.0 + - 7.0 COARSE: 75.0+- 7.0 TOTAL: 110.0+- 9.9
201
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CHB SITE: BACKGROUND YEAR: 81
FINE PARTICIPATE FRACTION
SAMPLING DURATION: 12 HRS. UITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RIDGE REGRESSION
K= .120 R-SQUARE:
CODE
SOURCE
INTERCEPT
1
2
3
5
6
9
10
11
13
14
16
17
19
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
DIST OIL
RES GAS
COAL
ELARCFRN
FERRMNFR
CEMENT
PET FCC
SEC S04
TOTAL:
SPECIES FIT
CODE
1
2
3
4
5
6
7
B
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
24
27
28
29
30
31
FLG
TOTAL *
AL *
AS *
BA *
BR *
CA *
CD *
CL *
CO *
CR
CU •*
FE *
HG
K *
HN *
NI *
P *
PB *
S04 *
SB *
SI *
SN
SR *
TI
V
ZN *
C
NA
N03
RB
SE
UG/M3
.006+-
-.334+-
.214+-
-.187+-
12.53B+-
1 .889+-
1.572*-
-.735+-
I.08B+-
-.053+-
.147+-
2.884+-
4.596+-
4.025*-
27.650+-
NISS
.006
.346
.654
.093
7.325
.609
3.786
.733
.951
.114
.137
1.166
4.504
4.481
7.882
FINE
FLG HEAS. UG/M3
28.000+-
.070+-
.014+-
.146+-
.164+-
.128+-
f
.010+-
.008+-
.010+-
.027+-
.120+-
<
.186+-
.022*-
.003+-
.103*-
.894+-
14.780*-
.008+-
.189+-
.024+-
.003+-
<
.038+-
N <
N <
H
M <
M
3.000
.055
.006
.018
.010
.008
.001
.007
.003
.005
.003
.008
.001
.010
.004
.002
.018
.047
.765
.003
.027
.003
.002
.008
.003
.003
.000
.000
.000
.000
.000
z
.622+- 1
1.171+- 2
.170+-
44.778I--26
6.745+- 2
6.768+-13
1.316+- 2
3.887+- 3
.204+-
.523+-
10.300+- 4
I6.416+-16
8.038+-16
98.750+-30
DATE: 0229
.8287
LF.AST
K-0.0
SQUARES
R-SQUARE:
UG/M3
.243
.343
.341
.596
.291
.536
.632
.420
.408
.492
.307
.183
.076
.073
-14
t
-
-26
2
20
17
3
-
9
6
7
27
.005+-
.654+-
.231+-
.667+-
.438+-
.261+-
.923+-
.660+-
.035+-
.117+-
.186+-
.390+-
.337+-
.995+-
.147+-
.006
34.854
3.814
.264
39.874
.522
1 2 . 055
61.095
2.132
.974
.498
3.855
24.170
30.701
20.851
7594
91
19
52
1
25
6725
13
93
22
30
55
78
.9241
VIF
.633
.330
.554
.511
.823
.512
.402
.675
.757
.554
.223
.903
.138
SUSPENDED PARTICULATE
PERCENT
100.000+-15.
.250+- .
.050+- .
.521+- .
.586+- .
.457+- .
<
.036+- .
.029+- .
.036+- .
.096+- .
.429+- .
<
.664+- .
.079+- .
.011+- .
.368+- .
3.193+- .
52.786+- 6.
.029+- .
.675+- .
.086*- .
.011+- .
<
<
.136+- .
^ ,
<
' .
c
<
152
198
022
085
072
057
007
025
011
018
015
054
OJ1
080
017
007
075
381
281
Oil
121
014
007
057
032
018
004
004
004
004
004
CALC. UG/M3
27.650+-
.048+-
.006+-
.006+-
.161 +-
.133+-
.007+-
.023+-
.006+-
.005+-
.011+-
.116+-
.006+-
.086+-
.029+-
.007+-
.006+-
.420+-
14.288+-
.006+-
.225+-
.015+-
.006+-
.008+-
.006+-
.035+-
6.858+-
-.038+-
.030+-
.006+-
.006+-
7
1
1
.882
.006
.000
.000
.039
.039
.000
.037
.000
.000
.001
.018
.000
.015
.006
.000
.000
.100
.728
.000
.029
.002
.000
.000
.000
.005
.259
.015
.004
.000
.000
RATIO
.988+-
.692+-
.403+-
.040+-
.983+-
1 .039+-
6.613+- 1
2.319+- 9
.705+-
.487+-
.426+-
.963+-
3.763+-
.461+-
1.303+-
2.463+-
.055+-
.470+-
.967+-
.745+-
1.188+-
.609+-
1.850+-
.978+-
1.925+-
.911+-
.000+-
.000+-
.000+-
.000+-
.000+-
.396
.110
.000
.000
.331
.437
.623
.227
.022
.031
.041
.205
.377
.089
.482
.421
.001
.123
.163
.012
.236
.109
.016
.086
.019
.181
.000
.000
.000
.000
.000
TOIAL
AL
AS
BA
6R
CA
CD
CL
CO
CR
CU
FE
HG
K
HN
NI
P
PB
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
NU3
RB
SE
KEASURED AMBIENT MASS (UG/M3): FINE: 28.0+- 3.0 COARSE: 37.0+- 4.0 TOTAL: 65.0+- 5.0
202
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CM» SITE: BACKGROUND YEAR: 81
COARSE PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. WITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: HO
RIDGE REGRESSION
K= .480 R-S0UARE:
CODE
SOURCE
INTERCEPT
!
2
3
5
&
9
10
11
12
13
14
14
17
19
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
DIST OIL
RES GAS
COAL
KRAFT RB
ELARCFRN
FERRMNFR
CEMENT
PET FCC
SEC S04
TOTAL:
SPECIES FIT
CODE
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FL6
TOTAL *
AL *
AS
BA *
BR *
CA *
CD
CL *
CO *
CR *
CU *
FE *
HG
K *
NN *
NI
P *
PB *
S04
SB
SI *
SN
SR «
TI *
V
ZN *
C
NA
N03
RB
SE
UG/H3
.003+-
20.253+-
6.130+-
-.024+-
1.749+-
.540+-
.000*-
.000+-
1 .397+-
.053+-
.000+-
.000+-
6.954+-
.000+-
.000+-
37.055+-
HISS
FLG HEAS.
37.000+-
./3V+-
..'
.086+-
.064+-
3.805+-
.014+-
<
<
.004+-
.692*-
.009*-
.215*-
.018*-
'
<
.213+-
.852*-
f
4.000+-
.003+-
. 004 + -
.097+-
<
.014+-
M <
h <
H <
H <
H <
.002
5.86!
1.363
.031
4.822
.104
.000
.000
.840
.538
.000
.000
2.494
.000
.000
13.502
COARSE
UG/M3
4.000
.171
.001
.016
.005
.444
,001
.010
.001
.003
.002
.044
.003
.032
.004
.001
.020
.017
.783
.001
.914
.003
.002
.017
-003
.003
.000
.000
.000
.000-
.000
z
54.738+-14
16.567+- 4
.042*-
4.522+-13
1.440+--
.000*-
.000+-
3.775+- 2
.726+- 1
.000+-
.000+-
I8.7?6*- 7
.000+-
.000+-
100.148+-38
DATE: 0229
.8785
LEAST
K=0.0
SQUARES
R-SOUARE:
UG/N3
.911
.097
.084
.043
.327
.000
.000
.307
.453
.000
.000
.558
.000
.000
.064
27
7
-
-9
2
-
9
38
.003+-
.211+-
.989+-
.053+-
.926*-
.820*-
.000*-
.000+-
.455*-
.154+-
.000+-
.000+-
.610+-
.000+-
.000+-
.153+-
.003
14.9)9
2.435
.114
14.544
.099
.000
.000
1.708
2.073
.000
.000
3.288
.000
.000
21,236
7
3
10
8
1
4
10
1
.9570
VIF
.651
.687
.065
.157
.130
.000
.000
.580
.701
.000
.000
.852
.000
.000
SUSPENDED PARTICULATE
PERCENT
100.000+-15.
1.997+- .
V. •
.232+- .
.173+- .
10.284+- 1 .
<
.038+- .
<
.011+- .
1.870+- .
.024+- .
.581+- .
.049+- .
<
<
.574+- .
2.303+- 2.
<
10.811+- 2.
.004+- .
.014+- .
.242+- .
<
.043+- .
C
<
<
s
< •
289
510
011
050
023
634
005
027
008
014
006
235
009
107
012
005
060
077
131
008
733
008
006
014
019
009
003
003
003
003
003
CALC. UG/K3
37.055+-13
.521+-
.003*-
.003+-
.048+-
2.444+-
.003+-
.020+-
.003+-
.003+-
.005+-
.513*-
.003+-
.167*-
.014*-
.003+-
.003+-
.121+-
1.315+-
.003+-
2.981+-
.003+-
.004+-
.083+-
.005*-
.014+-
2.6IO+-
.106+-
.004+-
.003+-
.003+-
.502
.105
.000
.000
.011
.436
.000
.008
.000
.000
.000
.092
.000
.035
.002
.000
.000
.029
.229
.000
.484
.000
.001
.014
.000
.002
.386
.027
.000
.000
.000
RATIO
1.001+-
./05+-
2.948+-
-.034+-
.746+-
.696+-
2.948+-
1.460+- 1
2.948+-
1 .281+-
1.311+-
.741+-
.328+-
.777+-
.757+-
3.327+-
.147+-
.548+-
1.544+-
2.948+-
.745*-
.983*-
.944+-
.854+-
1.315+-
.98?+-
.000+-
,000+-
.000+-
.000+-
.000+-
.516
.173
.000
.000
.216
.139
.000
.013
.697
.041
.189
.166
.026
.208
.130
.247
.011
.154
.495
.000
.151
.000
.144
.219
.197
.216
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CP
CL
CO
CR
CU
FE
HG
K
UN
NI
P
PB
SU4
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
HEASURED AHBIENT HASS (UG/M3): FINE: 28.0+- 3.0 COARSE: 37.0+- 4.0 TOTAL: 65.0+- 5.0
203
-------
TABLE C-7 (CONTINUED)
RESULTS FOR CMB SITE: BACKGROUND TEAK: 81 DAIE: 0229
TOTAL PARTICULATE FRACTION
SAMPLING DURATION: 12 HRS. WITH START HOUR: 7
BACKGROUND SITE SUBTRACTED: NO
RESULTS DERIVED FROM FINE AND COARSE FITTINGS
CODE
1
2
3
5
6
9
10
11
12
13
14
16
17
19
SOURCE UG/H3
INTERCEPT .009+-
SOIL 19.919+-
RD DUST 6.343+-
SEA SALT -.211+-
AUTO CAT 14.287+-
AUTO 2.429+-
DIST OIL 1.572+-
RES GAS -.735+-
COAL 2.485+-
KRAFT RB .053+-
ELARCFRN -.053+-
FERRhNFR .147+-
CEMENT 9.838+-
PET FCC 4.596+-
SEC S04 4.025+-
.006
5.872
1.513
.098
8.770
.618
3.786
.733
1.269
.538
.114
.137
2.935
4.504
4.481
TOTAL: 64.705+- 15.634
SPECIES FIT MISS TOTAL
CODE
1
2
3
4
5
6
7
8
9
10
It
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FL6 FLG MEAS. UG/M3
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
UN
HI
P
PB
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
N03
RB
SE
* 45.000+-
» .809+-
.015+-
« .232+-
* .228+-
* 3.933+-
<
* .024+-
* .009+-
* .012+-
* .031+-
* .812+-
.010+-
* .401+-
* .040+-
.004+-
* .123+-
* 1.I07+-
15.632+-
.009+-
* 4.189+-
.027+-
* .009+-
* .105+-
<
• .054+-
H
H <.
It
H <
M
5.000
.180
.007
.024
.011
.444
.002
.012
.004
.007
.004
.045
.004
.034
.006
.003
.028
.050
1.095
.004
.914
.004
.003
.023
.006
.004
.000
.000
.000
.000
.000
z
30.6431- 9.3J6
9.759+- 2.441.,
.077+- .153
21.980+-I3.597
3. 736+- .993
2.914+- 5.828
.565+- 1.131
3.823+- 1 .974
.413+- .827
.088+- .176
.225+- .211
15.136+- 4.663
7.07!+- 4.951
3.455+- 6.910
99.546 + -25.1M2
SUSPENDED PARTICULATE
PERCENT
100.000+-10
1.245+-
.023+-
.357+-
.351+-
4.051+-
<
.037+-
.014+-
.019+-
.048+-
1.249+-
.016+-
.617+-
.062+-
.006+-
.189+-
1 .703+-
24.049+- 2
.014+-
6.445+- t
.042+-
.014+-
.162+-
<
.083*-
(
<.
(
s
s
<
.879
.292
.011
.046
.032
.827
.004
.019
.007
.011
.007
.118
.007
.070
.010
.004
.046
.152
,502
.007
.492
.007
.004
.038
.018
.009
.002
.002
.002
.002
.002
CALC. U6/H3
64.705+-15
.569+-
.009+-
.009+-
.209+-
2.780+-
.010+-
.044+-
.009+-
.008+-
.017+-
.628+-
.009+-
.253+-
.042+-
.011+-
.009+-
.541+-
15.603+- 1
.009+-
3.205+-
.018+-
.011+-
.091+-
.010+-
.050+-
9.468+- 1
.067+-
.034+-
.009+-
.009+-
.634
.105
.000
.000
.040
.437
.000
.037
.000
.000
.001
.094
.000
.038
.007
.000
.000
.104
.743
.000
.485
.002
.001
.016
.000
.006
.317
.031
.004
.000
.000
RATIO
.995+-
.704+-
.573+-
.038+-
.916+-
.707+-
4.780+-
1 .818+-
^S1:)*-
.646+-
.540+-
.774+-
.813+-
.631+-
1.057+-
2.679+-
.070+-
.489+-
.998+-
.990+-
.765+-
.650+-
1.259+-
.864+-
1.596+-
" .934+-
.000+-
.000+-
.000+-
.000+-
.000+-
.339
.159
.000
.000
.240
.136
.593
3.233
.041
.027
.041
.146
.033
.113
.245
.343
.002
.104
.158
.012
.146
.099
.127
.203
.121
.143
.000
.000
.000
.000
.000
TOTAL
AL
AS
BA
BR
CA
CD
CL
CO
CR
CU
FE
HG
K
MN
NI
P
PB
S04
SB
SI
SN
SR
TI
V
ZN
C
NA
NO 3
RB
SE
AMPIFNT HAS; 'Mr,, HI):
70.0 + -
CW5F : V.0'~
204
-------
TABLE C-7 (CONTINUED)
RIDGE K
INTERCEPT
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
COAL
KRAFT RB
CEMENT
RID6E K
INTERCEPT
SOIL
ID DUST
SEA SALT
AUTO CAT
AUTO
COAL
KRAFT RB
CEhENT
RID6E K
INTERCEPT
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
COAL
KRAFT RB
CEMENT
RIDGE X
INTERCEPT
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
COAL
KRAFT RB
CEHENT
RIDSE K
INTERCEPT
SOIL
RD DUST
SEA SALT
AUTO CAT
AUTO
COAL
KRAFT RB
CEMENT
.000
.004*-
22.079+-19
8.616+- 3
-.034*-
-6.547+-16
.847*-
2.26B+- 2
-.203+- 1
50.305+- 5
.035
.003+-
18.106+-12
8.730*- 2
-.043+-
-2.310+-10
.810*-
1.805*- 1
-.270*- 1
10.399+- 4
.090
.003*-
14.159*- 8
8.494*- 2
-.037*-
.501*- 4
.758+-
1.486*- 1
-.267+- 1
10.035+- 3
.240
.003+-
15.17B+- 5
7.624+- 2
-.027+-
3.143*- 4
.447*-
1.170*-
-.181+-
8.910+- 3
.640
.002+-
14.575+- 4
5.967+- 1
-.013+-
4.717+- 3
.476+-
.945+-
-.023+-
6.983+- 2
.003
.624
.282
.120
.349
.286
.178
.980
.092
.002
.136
.739
.091
.177
.263
.492
.506
.411
.002
.252
,448
.066
.792
.236
.128
.105
.870
.002
.531
.082
.041
.413
.192
.865
.686
.17?
.002
.431
.746
.028
.532
.146
.736
.473
.636
.005
.004+- .003
21.180+-17.940
8.647+- 3.164
-.032*- .115
-5.666+-15.120
.841*- .282
2.174+- 2.022
-.220+- 1.894
10.345*- 4.944
.040
.003*- .002
17.804*-11.545
8.720*- 2.719
-.044*- .088
-1.936+- 9.684
.805+- .260
1.743+- 1.439
-.272+- 1.457
10.37B+- 4.346
.100
.003+- .002
15.993+- 7.878
8.437+- 2.411
-.037*- .063
.817+- 6.462
.749+- .232
1.449+- 1.092
-.242+- 1.055
9.935+- 3.799
.280
.003+- .002
15.10?*- 5.241
7.414*- 2.024
-.025+- .037
3.474+- 4.185
.424+- .184
1.128+- .839
-.159*- .436
8. 655+- 3.076
.720
.002+- .002
14.452*- 4.380
3.725*- 1.708
-.013*- .027
4.811*- 3.300
.454*- .141
.725*- .730
-.003*- .462
4.710*- 2.384
.010
.004+- .002
20.448+-16.547
8.701+- 3.066
-.050+- .110
-4.919+-13.942
.834+- .278
2.093+- 1.896
-.234+- 1.815
10.399+- 4.842
.045
.003+- .002
17.542+-I1.057
8.704+- 2.6B2
-.043+- .085
-1.594+- f.246
.800+- .257
1.723+- 1.392
-.274+- 1.411
10.353+- 4.284
.120
.003*- .002
15.742+- 7.274
8.320+- 2.344
-.035+- .038
1.331+- 5.930
.732*- .225
1.387+- 1.034
-.252+- .971
9.794+- 3.672
.320
.003+- .002
15.054*- 5.05?
7.218*- 1.976
-.024*- .035
3.738+- 4.018
.602+- .178
1.094*- .81?
-.139*- .598
8.419+- 2.992
.800
.002+- .002
14.300+- 4.343
5.502+- 1.674
-.012+- .026
4.875+- 3.47?
.433*- .137
.708+- .744
.014+- .454
6.460+- 2.542
.015
.004+- .002
19.828+-15.377
B. 722+- 2.984
-.049+- .105
-4.267+-12.749
.831*- .275
2.02I+- 1.790
-.245+- 1.743
10.419+- 4.735
.050
.003+- .002
17.309+-10.608
8. 689+- 2.648
-.042+- .082
-1.284+- 8.834
.795+- .235
1.490+- 1.350
-.275+- 1.368
10.323*- 4.22?
.140
.003*- .002
15.367+- 6.810
8.200+- 2.286
-.033+- .054
1.785+- 3.522
.716+- .218
I.333+- .990
-.240+- .901
9.435*- 3.542
.360
.003+- .002
15.004+- 4.905
7.030+- 1.934
-.022+- .033
3.952*- 3.892
.582*- .172
1.065+- .804
-.120+- .368
8.199+- 2.921
.020
.004+-
.002
1?.27?+-14.382
8.734*-
-.048+-
2.716
.101
-3.692*-12.102
.825+-
1.95B+-
-.254*-
10.427*-
.060
.003+-
16.920+-
8.648*-
-.041*-
-.733*-
.785+-
1.627+-
-.276*-
10.260*-
.160
.003+-
15.441+-
8.081*-
-.032*-
2.147*-
.701+-
1.292+-
-.228*-
7.480*-
.400
.003*-
14.955+-
6.853+-
-.021+-
4.128*-
.544+-
1.040+-
-.103+-
7.793+-
.272
1.679
1.677
4.641
.002
9.841
2.588
.078
8.184
.250
1.278
1.290
4.125
.002
6.443
2.235
.050
5.201
.212
.954
.843
3.463
.002
4.785
1.878
.032
3.797
.167
.792
.545
2.862
.025
.004+-
18.844*-
8.73?*-
-.047*-
-3.I81+-
.81'0 + -
1.902*-
-.260*-
10.425*-
.070
.003+-
14.60?+-
8.401*-
-.040+-
-.244+-
.776*-
1.574*-
-.274*-
10.189*-
.180
.003+-
15.349*-
7.963*-
-.031+-
2.453+-
.687*-
1.235*-
-.214*-
9.329+-
.480
.003+-
14.845*-
6.527*-
-.018+-
4.3V7+-
.331+-
1.001*-
-.072*-
7.417*-
.002
13.527
2.857
.098
11 .371
.249
1.420
1.616
4.557
.002
7.214
2.536
.073
7.637
.245
1.21?
1.221
4.032
.002
6.147
2.190
.047
4.943
.206
.925
.794
3.380
.002
4.615
1.838
.030
3.664
.159
.775
.511
2.769
.030
.004*- .
18.449+-12.
8.737*- 2.
-.046*- .
-2.723+-10.
.815+- .
1.851+- 1.
-.266+- 1.
10.413+- 4.
.080
.003*- .
16.360*- 8.
8.549+- 2.
-.039+- .
.144*- 7.
.767*- .
1.527*- 1.
-.271+- 1.
10.113+- 3.
.200
.003+- .
15.279+- 5.
7.848*- 2.
-.029+- .
2.713*- 4.
.673*- .
1.22,!+- .
-.204+- .
9.1B4+- 3.
.560
.002*- .
14.72;'*- 4.
6.234*- 1.
-.016+- .
4.584+- 3.
.502*- .
.970*- .
-.046+- .
7.28«+- 2.
002
783
805
094
733
266
532
337
481
002
692
490
049
178
240
170
140
948
002
904
150
043
733
201
902
752
305
002
505
788
029
583
152
764
488
673
205
-------
TABLE C-8. USER FOLLOWUP OUTPUT FILE
3
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
3
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
?
7
7
7
1
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
/
7
7
7
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
35.0000
.0880
.0360
.1490
.3440
.4410
.0040
.0510
.0030
.0020
.0130
.2080
.0000
.1650
.0220
.0060
.1190
1 .4220
13.1300
.0030
.2770
.0010
.0030
.0070
.0020
.0420
35.0000
.0880
.0360
.1490
.3440
.4410
.0040
.0510
.0030
.0020
.0130
.2080
.0000
.1650
.0220
.0060
.1190
7.0000
.0620
.0070
.0140
.0170
.0230
.0020
.0070
.0020
.0040
.0020
.0110
.0030
.0090
.0040
.0010
.0170
.0680
.6810
.0020
.0310
.0020
.0010
.0140
.0080
.0030
7.0000
.0620
.0070
.0140
.0170
.0230
.0020
.0070
.0020
.0040
.0020
.0110
.0030
.0090
.0040
.0010
.0170
75.0000
1.9510
.0040
.3710
.1080
13.0140
.0005
.2880
.0150
.0220
.0180
1.8710
.0010
.5170
.0880
.0030
.1020
.3460
1.7250
.0040
8.6520
.0050
.0300
.2050
.0070
.0710
75.0000
1.9510
.0040
.3710
.1080
13.0140
.0005
.2880
.0150
.0220
.0180
1.8710
.0010
.5170
.0880
.0030
.1020
7.0000
.4220
.0040
.0230
.0070
1.4420
.0010
.0680
.0030
.0050
.0020
.1180
.0020
.0720
.0080
.0010
.0400
.0240
.9390
.0020
1.9870
.0020
.0030
.0370
.0140
.0060
7.0000
.4220
.0040
.0230
.0070
1.4420
.0010
.0680
.0030
.0050
.0020
.1 180
.0020
.0720
.0080
.0010
.0400
206
-------
TABLE C-8 (CONTINUED)
,50
30
30
,30
,50
30
,50
30
30
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
810229
310229
810229
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
HI 022 9
810229
81022?
810229
810229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
1
7
7
7
7
7
7
7
7
7
7
7
7
?
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
->
/
->
/
7
7
7
18
19
20
21
22
23
24
25
26
1 1
1 2
1 6
1 11
1 12
1 14
1 15
1 18
1 19
1 21
1 23
1 24
1 26
1 30
2 1
2 2
2 5
2 6
2 10
2 11
2 12
2 14
2 15
2 16
2 18
2 19
2 21
2 24
2 25
2 26
2 27
2 28
2 29
3 1
3 6
3 8
3 14
3 19
1 .4^20
13.1300
.0030
.2770
.0010
.0030
.0070
.0020
.0420
-.5088
-.0033
-.0473
-.0002
-.0042
-.0010
-.0002
-.0001
-.0023
-.0122
-.0001
-.0007
-.0001
.0000
-.0244
-.0022
.0000
-.0006
.0000
.0000
-.001'!!/
-.0003
.0000
.0000
-.0001
-.0003
-.0054
-.0002
.0000
.0000
-.0033
-.0003
.0000
-.4464
-.0045
-.2455
-.0045
-.0402
.0680
.6310
.0020
.0310
.0020
.0010
.0140
.0080
.0030
3.0917
.0207
.2966
.0010
.0265
.0064
.0012
.0004
.0144
.0765
.0009
.0045
.0007
.0001
.9225
.0841
.0002
.0232
.0004
.0003
.0571
.0098
.0012
.0001
.0035
.0106
.2121
.0061
.0002
.0010
.1298
.0119
.0000
.1712
.0021
.1148
.0021
.0188
..5460
1.7250
.0040
8.6520 '
.0050
.0300
.2050
.0070
.0710
36.4449
.2004
1.8587
.0033
.1749
.0437
.0077
.0073
.0984
.8747
.0051
.0324
.0033
.0007
13.5780
.8948
.0011
.4073
.0000
.0000
.7780
.0000
.0136
.0000
.0000
.0000
3.8561
.1371
.0037
.0000
.6640
.2376
.0027
.1769
.0018
.0975
.0013
.0157
.OJ40
.9390
.0020
1 .9870
.0020
.0030
.0370
.0140
.0060
18.7516
.1175
1 .0898
.0019
.1026
.0256
.0045
.0043
.0577
.5128
.0030
.0190
.0019
.0004
6.1852
.4760
.0006
.2167
.0000
.0000
.4139
.0000
.0072
.0000
.0000
.0000
2.0513
.0730
.0020
.0000
.3532
.1264
.0014
.4910
.0051
.2794
.00'51
.045;'
207
-------
TABLE C~8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
8.10229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
810229
81v22V
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
-i
/
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
/
1
7
7
7
7
7
7
7
7
7
7
7
7
7
7
1
7
7
7
3
5
5
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
9
9
9
9
9
9
9
9
9
9
9
9
9
7
9
9
9
9
10
10
10
28
1
2
6
11
12
14
15
16
19
21
26
27
1
2
5
8
11
12
18
19
21
26
27
1
2
5
6
8
11
12
14
15
16
18
19
21
25
26
27
28
29
1
6
19
-.1384
7.3772
.0089
.0125
.0018
.0081
.0032
.0011
.0011
3.6386
.0376
.0059
2.8771
4.6355
.0020
.3801
.2503
.0002
.0116
.9781
.0097
.0035
.0010
2.5263
3.5958
.0111
.0009
.0180
.0431
.0061
.0043
.0006
.0005
.0003
.0194
.4746
.0097
.0002
.0010
1.2873
.0115
.0360
1.0147
.0507
.476V
.0647
20.4363
.0254
.0359
.0051
.0233
.0093
.0032
.0032
10.5729
.1078
.0169
8.2469
.9405
.0007
. 1 239
.0816
.0001
.0038
.3188
.0032
.0011
.0003
pOTIS
. 0,^0 j
6.6142
.0213
.0018
.0344
.0825
.0117
.0083
.0012
.0010
.0006
.0371
.9077
.0186
.0003
.0020
2.461V
.0220
.0688
5.3113
.2740
2.5759
.0548
10.3394
.0124
.0176
.0025
.0114
.0045
.0016
.0016
5.1697
.0527
.0033
4.0324
.7224
.0003
.0592
.0390
.0000
.0018
.1524
.0015
.0005
.0002
.3937
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.1575
13.6553
.0172
.0243
.0034
.0157
.0063
.0021
.0021
7.1555
.0730
.0114
5.5813
.2752
.0001
.0276
.0182
.0000
.0008
.0710
.0007
.0003
.0001
.1833
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
208
-------
TABLE C-8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
LORE
810229
810229
310229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
81022?
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
10
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
13
13
13
13
13
13
13
13
13
27
1
2
4
6
7
10
11
12
14
16
18
19
20
21
22
24
26
28
1
2
5
6
8
10
11
12
14
15
16
19
21
27
28
1
2
6
8
10
11
12
14
1':i
.1218
1.7708
.0078
.0002
.0322
.0016
.0003
.0004
.0767
.0033
.0001
.0118
.1705
.0005
.0630
.0146
.0021
.0288
.0043
. 0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0763
.0005
.0047
.0014
.0016
.0002
.0244
.0007
.0066
.6577
1.3477
.0064
.0002
.0265
.0013
.0002
.0003
.0632
.0028
.0001
.0098
.1404
.0004
.0518
.0120
.0018
.0237
.0035
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.4535
.0030
.0290
.0087
.0098
.0013
.1495
.0043
.0407
.0000
4.5322
.0000
.0000
.1070
.0000
.0000
.0000
.1858
.0000
.0000
.0000
.5489
.0000
.1492
.0000
.0000
.0310
.0000
2.2289
.0062
.0012
.0080
.0646
.0107
.0013
.0410
.0089
.0012
.0049
.2630
.0029
.3923
.1181
.0000
.0000
.0000
. 0000
.0000
.0000
.0000
.0000
.0000
.0000
3.1369
.0000
.0000
.0809
.0000
.0000
.0000
.1404
.0000
.0000
.0000
.4150
.0000
.1128
.0000
.0000
.0234
.0000
3.4552
.0101
.0020
.0130
.1045
.0173
.0022
.0663
.0144
.0019
.0079
.4254
.0047
.6345
,1911
.0000
.0000
.0000
.0000
.0000
.0000
. 0000
.0000
.0000
209
-------
TABLE C-8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
3
,50
30
30
30
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
URBAN
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
CORE
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
81022?
810229
810229
81022?
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
81022?
810229
810229
810229
810229
810229
810229
81022?
810229
81022?
810229
810229
810229
810229
810229
3102129
81022?
81022?
810229
810229
310229
810229
810229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
/
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
•J
/
7
7
7
7
7
7
7
/
13
13
13
13
13
13
13
13
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
U
U
16
16
16
16
17
17
17
17
19
19
16
18
19
21
24
25
26
28
1
2
5
6
8
10
11
12
14
15
18
19
21
24
25
26
27
28
2?
1
6
14
19
21
27
1
12
1?
27
1
1?
12
1
n
3
4
.0005
.0006
.001?
.0038
.0002
.0000
.0009
.0010
.0605
.0004
.0001
.0008
.0003
.0000
.0000
.0013
.0064
.0105
.0000
.0025
.0006
.0000
.0000
.0004
.0064
.0019
.0034
7.4027
.3701
.1481
4.4416
.1851
.5922
8.2208
.0822
4.1104
.3288
-.1872
-.1872
28.0000
.0700
.0140
.1460
.0033
.0036
.0117
.0234
.0009
.0003
.0056
.0059
.2501
.0017
.0004
.0034
.0011
.0001
.0001
.0054
.0271
.0447
.0001
.0108
.0026
.0001
.0001
.0015
.0271
.0080
.0147
2.9194
.1766
.0707
2.1197
.0883
.2826
12.2522
.1280
6.3977
.5118
11.0124
11.0124
3.0000
.0550
.0060
.0180
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
13.8741
2.7743
.2775
.4162
1 .3874
2.6361
.0000
.0000
.0000
.0000
.0000
.0000
37.0000
.7390
.0010
.0860
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
11.3240
2.4354
.2435
.3653
1 .2177
2.3136
.0000
.0000
.0000
.0000
.0000
.0000
4.0000
.1710
.0040
.0160
210
-------
TABLE C-8 (CONTINUED)
30 BACKGROUND
JO BACKGROUND
,50
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
'10
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
810229 12
810229 12
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
1
7
7
7
7
7
7
7
7
1
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
->
i
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
T
T
K
•J
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1
2
6
11
12
14
15
18
19
21
23
24
26
30
1
2
5
6
10
11
12
14
15
16
18
19
14
.1640
.1280
.0010
.0100
.0080
.0100
.0270
.1200
.0015
.1360
.0220
.0030
.1030
.8940
.7800
.0080
.1890
.0240
.0030
.0080
.0030
.0380
-.3337
-.0022
-.0310
-.0001
-.0028
-.0007
-.0001
.0000
-.0015
-.0080
•.0001
-.0005
-.0001
.0000
.2138
.0189
.0000
.0052
.0001
.0001
.0128
.0022
.0003
.0000
.0008
.0024
.0100
.0080
.0020
.0070
.0030
.0050
.0030
.0080
.0030
.0100
.0040
.0020
.0180
.0470
.7650
.0030
.0270
.0030
.0020
.0160
.0090
.0030
.3463
.0024
.0341
.0001
.0030
.0007
.0001
.0001
.0016
.0088
.0001
.0005
.0001
.0000
.6555
.0599
.0001
.0165
.0003
.0002
.0407
.0070
.0008
.0001
.0025
.0075
20
.0640
.8050
.0010
.0140
,0010
,0025
,0040
,6920
,0090
,2150
,0180
,0010
0200
2130
8520
0010
0000
0030
0060
0970
0035
0160
2530
.1114
I.0329
.0018
.0972
.0243
.0043
.0041
.0547 '
.4861
.0028
.0180
.0018
.0004
. 1 296
.4039
.0005
.1839
.0000
.0000
.3512
.0000
.0061
.0000
.0000
.0000
.0050
.4440
.0020
.0100
.0030
.0050
.0020
.0440
.0030
.0320
.0040
.0020
.0220
.0170
.7830
.0030
.9140
.0030
.0020
.0170
.0070
.0030
3.8613
.0434
.4020
.0007
.0378
.0095
.0017
.0016
.0213
.1392
.0011
.0070
.0007
.0002
.3633
.1370
.0002
.0624
.0000
.0000
.1191
.0000
.0021
.0000
.0000
.0000
211
-------
TABLE C-8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
H10229
810229
HI 0229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
"7
/
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
2
2
2
2
2
2
2
3
3
3
3
3
3
5
5
5
5
5
5
5
5
5
5
5
5
6
6
' 6
6
6
6
6
6
6
6
6
9
9
9
V
9
9
21
24
25
26
27
28
29
1
6
8
14
19
28
1
2
6
11
12
14
15
16
19
21
26
27
1
2
5
8
11
12
18
19
21
26
27
1
2
5
6
8
1 1
.0477
.0014
.0000
.0002
.0292
.0027
.0000
-.1872
-.0019
-.1029
-.0019
-.0168
-.0580
12.5377
.0150
.0213
.0030
.0138
.0055
.0019
.0019
6.2689
.0639
.0100
4.8897
1.8886
.0008
.1'549
.1020
.0001
.0047
.3935
.0040
.0014
.0004
1 .0293
1 .5719
.0049
.0004
.0079
.0189
.0027
.1512
.0043
.0002
.0007
.0925
.0085
.0000
.0933
.0011
.0538
.0011
.0096
.0332
7.3249
.0098
.0139
.0020
.0090
.0036
.0012
.0012
4.0875
.0417
.0065
3.1883
.6088
.0003
.0644
.0424
.0000
.0020
. 1 657
.0016
.0006
.0002
.4280
3.7865
.0122
.0010
.0196
.0471
.0067
1 ./'408
.0619
.0017
.0000
.2997
.1073
.0012
-.0240
-.0002
-.0132
-.0002
-.0022
-.0074
1 .7490
.0021
.0030
.0004
.0019
.0008
.0003
.0003
.8745
.0089
.0014
.6321
.5401
.0002
.0443
.0292
.0000
.0014
.1140
.0011
.0004
.0001
.2943
.0000
. 0000
. 0000
.0000
. 0000
.0000
.5905
.0210
.0006
.0000
.1017
.0364
.0004
.0311
.0003
.0179
.0003
.0029
.0101
4.8223
.0060
.0085
.0012
.0055
.0022
.0007
.0007
2.4949
.0254
.0040
1.9461
.1058
.0001
.0142
.0094
.0000
.0004
.0366
.0004
.0001
.0000
.0946
.0000
.0000
.0000
.0000
. 0000
.0000
212
-------
TABLE C-8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKbROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
310229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
81022V
810229
810229
810229
810229
810229
810229
810229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
9
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
11
11
11
11
1
1
1
1
1
1
1
1
1
1
11
11
11
11
12
12
12
12
12
12
12
12
12
12
12
12
12
14
15
16
18
19
21
25
26
27
28
29
1
6
19
27
1
2
4
6
7
10
11
12
14
16
18
19
20
21
22
24
26
28
1
2
5
6
8
10
11
12
14
15
14
19
.0019
.0003
.0002
.0001
.0085
.2075
.0042
.0001
.0005
.5*28
.0050
.0157
-.7351
-.0368
-.3455
-.0882
1 .0883
.0048
.0001
.0198
.0010
.0002
.0002
.0472
.0021
.0001
.0073
.1048
.0003
.0387
.0090
.0013
.0177
.0026
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0047
.0007
.0005
.0004
.0212
.5178
.0106
.0002
.0011
1.4043
.0126
.0392
.7328
.0389
.3654
.0933
.9506
.0045
.0001
.0185
.0009
.0002
.0002
.0441
.0019
.0001
.0068
.0979
.0003
.0362
.0084
.0012
.0165
.0025
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
. 0000
.0000
.0000
. 0000
.0000
.0000
.0000
. 0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
1.3968
.0000
.0000
.0330
.0000
.0000
.0000
.0573
.0000
.0000
.0000
.1692
.0000
.0460
.0000
.0000
.0095
.0000
.0531
.0001
.0000
.0002
.0015
.0003
.0000
.0010
.0002
.0000
.0001
.0063
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.8400
.0000
.0000
.0220
.0000
.0000
.0000
.0383
.0000
.0000
.0000
.1131
.0000
.0307
.0000
.0000
.0064
.0000
.'53/5
.0016
.0003
.0020
.0161
.0027
.0003
.0102
.0022
.0003
.0012
.O6'j4
213
-------
TABLE C-8 (CONTINUED)
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
40
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
BACKGROUND
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
810229
81 0229
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
?
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
/
•>
7
7
7
7
7
1
7
7
7
~t
f
7
7
7
7
7
7
12
12
12
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
16
16
16
16
16
21
27
28
1
2
6
8
10
11
12
14
15
16
18
19
21
24
25
26
28
1
2
5
6
8
10
11
12 '
14
15
18
19
21
24
25
26
27
28
29
1
6
14
19
21
.0000
.0000
.0000
-.0525
-.0003
-.0033
-.0010
-.0011
-.0001
-.0168
-.0005
-.0046
-.0004
-.0004
-.0013
-.0026
-.0001
.0000
-.0006
-.0007
.1465
.0009
.0002
.0019
.0006
.0001
.0001
.0031
.0154
.0254
.0001
.0062
.0015
.0001
.0000
.0008
.0154
.0045
.0084
2.8841
.1442
.0577
1.7305
.0721
.0000
.0000
.0000
.1141
.0008
.0073
.0022
.0025
.0003
.0378
.0011
.0103
.0008
.0009
.0030
.0059
.0002
.0001
.0014
.0015
.1368
.0009
.0002
.0019
.0006
.0001
.0001
.0031
.0153
.0252
.0001
.0061
.0014
.0001
.0000
.0008
.0153
.0045
.0083
1.1653
.0701
.0280
.8403
.0350
.0001
.0093
.0028
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
6.9544
1.3909
.1391
.?086
.6954
.0007
.0975
.0294
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
2.6936
.6552
.0655
.0983
. J2/6
214
-------
TABLE C-8 (CONTINUED)
40 BACKGROUND
40 BACKGROUND
40 BACKGROUND
40 BACKGROUND
40 BACKGROUND
40 BACKGROUND
40 BACKGROUND
810229
810229
810229
810229
810229
810229
810229
12
12
12
12
12
12
12
7
7
7
7
7
7
7
16
17
17
17
17
19
19
27
1
12
19
27
1
19
4
2
4
4
.2307
.5964
.0460
.2982
.1839
.0254
.0254
•
4.
B
2.
m
4.
4.
1121
5043
0478
3915
1913
4807
4807
1.3213
.0000
.0000
.0000
.0000
.0000
.0000
.6224
.0000
.0000
.0000
.0000
.0000
.0000
215
-------
APPENDIX D
STATISTICAL APPROACH FOR SOURCE APPORTIONMENT
This appendix presents a discussion of the statistical methods used in
the source apportionment software documented herein. First, the source
apportionment problem is formulated in mathematical terms. This involves
primarily stating the mass balance equations used in source apportionment.
Subsequently a discussion of the statistical methods used in the soft-
ware documented herein is presented. The basic approach used is regression
analysis. However, several additional features are available to handle
specific aspects of the problem.
While Section 2 discusses source apportionment strictly from a user's
viewpoint, this appendix discusses the mathematics, with references to
papers which provide further detail. For the convenience of the reader,
Section 2 and this appendix have both been made as nearly self-contained
as is feasible. As a result, some overlap exists between the two discussions,
Mathematical Formulation of the Problem
The standard equation used in CMB analysis is as follows:
m
C. = I F..S. (1)
where
C. is the ambient particulate concentration of species i,
216
-------
F.. is the fraction of the particulate matter emitted from source j
comprised of species i,
S. is the ambient particulate concentration resulting from source j,
and
m is the number of sources.
All inputs pertain to the size fraction being considered.
The C. values are obtained through chemical analysis of ambient par-
ticulate samples. The F,. values are obtained through direct source
characterization or by indirect methods, such as sampling upwind and down-
wind of a source. The C. and F.. values are required inputs. Depending on
the type of statistical analysis performed, the standard errors of the C.
and F.. values may also be included. The "standard error" of a quantity is
the standard deviation of its random error. The values of S. represent the
contributions of the different sources to the ambient particulate concentra-
tion, and, as such, are the unknown values.
If multiple particulate samples have been collected for a specific
set of conditions, averaging can be performed to produce a single concen-
tration for each species. Subsequently, a single CMB analysis can be
performed using the averaged concentrations. An alternative is to perform
a separate CMB analysis for each particulate sample and then average the
CMB results for each source. In this case, however, the error present in
the source signature (F ) matrix would be common to all periods. Thus,
the separate CMB analyses would have correlated errors. For this reason,
the conventional standard error of the mean would have a low bias (see
the discussion of this point in Section 2).
At the option of the user, an intercept term S can be added to the
standard equation:
m
C. - S + I F..S. (2)
i o . , 11 j '
3=1
217
-------
line intercept represents a part of the particulate concentration not accoun-
ted for by the m sources considered. Thus, the intercept should be zero
within random error unless there is a significant source which has not been
considered or large data errors exist. The intercept provides a check on
the closure of the mass balance equations. If the regression is performed
with no intercept, all variances are computed about zero, not about the
mean values.
An additional equation can also be employed:
m
CT = SQ + ^ S. (3)
where C is the measured aerosol mass concentration for the size fraction
being considered. This equation simply states that the sum of the intercept
and the particulate concentrations attributable to the m sources equals the
ambient aerosol mass concentration. The intercept here is again the
particulate concentration not explainable by any of the sources considered.
Thus, as before, the intercept should be near zero unless there is a sig-
nificant omitted source or large data errors.
Statistical Approaches
Regression analysis is the basic statistical approach which has been em-
ployed in the program documented herein (Cooper et al., 1979; Kowalczyk et
al., 1978). Regression analysis is very versatile in that it can be modified
in various ways to handle the different conditions which can arise. Several
such properties of the CMB problem, which require special features not pro-
vided by ordinary multiple regression analysis, are discussed below.
First, in the regression analysis, the C. and C are the values of the
dependent variable, and it is known that these values have different error
variances. To account for this fact, weighted least squares should be used.
218
-------
This approach differs from ordinary multiple regression in that the most
accurate data are weighted most heavily in developing the regression
equation. In ordinary multiple regression, all data are weighted equally.
In weighted regression, the values of S and S., j=l to m, are chosen so as
to minimize the sum of squares SS:
Tl f) f,
f r* f ^ ^ f P f ^ ^
SS - Z \ i " i; + C T " V (4)
i=l~~il ~1I
Ci S
where
C. and C are concentrations predicted by equations (2) and (3),
respectively,
C. and C are the corresponding measured concentrations,
s?, and s?, are the error variances of C. and C , respectively, and
Ci CT il
n is the number of chemical species being considered.
Second, in ordinary multiple regression analysis, it is assumed that
the predictor variables have no error. In the CMB analysis, however, the
values of the predictor variables (F..) have random errors. Approaches to
account for these errors in the CMB analysis have been described by Watson
(1979). In one approach, the C. and F. . values are all weighted according
to their uncertainties in achieving a maximum-likelihood solution of the
mass balance equations (Eq. 1). The approach requires an iterative solution,
however, which sometimes fails to achieve convergence. A simpler method uses
"effective variances" to account for the errors in the F. . . In this approach,
the variance of the total error in each mass balance equation, including the
error of C. and the errors in the terms involving the F. ., is taken into
i~H
account. The resulting effective variance for the i equation is
m
=
-------
A weighted least squares analysis is performed, and effective variances are
used in place of error variances of ambient concentration measurement error
variances.
In the mass balance equation involving aerosol mass concentration (Eq.
3), the coefficients of the S. are all unity and have zero error variances.
Thus, the effective variance of this equation is simply the error variance
of C_. Using effective variances, then, increases the error variances
associated with all other equations (Eq. 2) relative to the error variance
for the equation involving aerosol mass concentration (Eq. 3). Thus, the
use of effective variances increases the relative weight given to the
aerosol mass equation (Eq. 3). Use of effective variances will not always
improve the agreement between the measured and predicted aerosol mass
concentrations, however. This is because of the complex shifting of the
relative weights given to all the equations.
Since the effective-variance calculation involves the unknown S. values,
an initial solution estimate is required. Thus, an iterative approach can
be used; the initial estimate at each stage is the solution from the preced-
ing iteration.
Another important aspect of the CMB approach involves the difficulty
in estimating the separate particulate contributions of sources whose
aerosols have similar chemical compositions. If the compositions were
identical, it would be impossible to differentiate between the aerosols
from the two sources. If the compositions were very similar, differentia-
tion would be possible but difficult to accomplish accurately. The problem
is referred to in statistical terms as "multicollinearity." Stated more
generally, multicollinearity exists when any source signature is very
nearly a linear combination of any subset of the other signatures. In this
case, it is difficult to distinguish among the aerosols emanating from the
sources with nearly linearly dependent signatures. Multicollinearity can
cause large random errors if conventional weighted least squares with or
220
-------
without effective variances is used. Negative values with large magnitudes
can result from these errors.
Ridge regression is an approach for handling problems with multi-
collinearity. A brief summary of the properties of ridge regression which
are essential to the discussion here is presented below. Ridge regression
is discussed further by Vinod (1978), Hoerl and Kennard (1970a, b),
Marquardt (1970), Swindel (1974), Marquart and Snee (1975), and Obenchain
(1977) . The use of ridge regression in source apportionment is discussed by
Williamson, Balfour, and Schmidt (1982).
In ordinary regression analysis,
B = (x'x^x'y (6)
where
B is the vector of regression coefficients - in CMB analysis, the inter-
cept S and the S.,
x is the design matrix, here the source signature matrix,
y is the vector of values of the dependent variable, here the species
and total.particulate concentrations,
the prime indicates matrix transposition,
and the superscript ~ indicates matrix inversion.
In ridge regression, a parameter k is introduced:
B(k) = (x'x + kl) x'y (7)
where
B(k) is the ridge solution corresponding to the parameter k, and
I is the identity matrix.
Notice that if k = 0, the ridge solution becomes the ordinary regression
solution.
221
-------
The usual practice is to standardize x'x and y.'y to contain correlation
coefficients. The B values obtained from Eq. 6 are then called standardized
coefficients. The solution in terms of the original data units is obtained
by an inverse transformation. The diagonal elements of x'x, then, are all
one. A k value of, for example, 0.1 represents a 10 percent perturbation of
the diagonal elements. If k is positive, a bias is introduced into the
regression coefficients; the conventional regression solution is unbiased.
However, the error variances of the coefficients, if inflated due to multi-
collinearities, often decrease rapidly as k increases. Thus, a much improved
solution can be achieved when, by introducing a small bias, a large reduction
of the random errors is accomplished.
Variance-inflation factors (Belsley et al., 1980) are convenient mea-
sures of the effect of multicollinearity in a given problem. The variance-
inflation factor of a given regression coefficient (S.) can be interpreted
as the ratio of its error variance to the error variance it would have had
if the predictor variables had been uncorrelated.
The covariance matrix C of the standardized regression coefficients is
C = (x'x)'1 a2 (8)
where a2 is the error variance of the dependent variable. The error variances
of the regression coefficients are the diagonal elements of C. The error
variances which would have been obtained if the predictor variables were un-
correlated are obtained by setting the off-diagonal elements of x'x equal to
zero and employing Eq. 8. The variance inflation factors, then, are the
ratios of the actual error variances to those obtained assuming uncorrelated
predictor variables. The variance inflation factors apply either to the
standardized or unstandardized regression coefficients, since the scale
factors cancel in taking the ratio.
222
-------
If multicollinearities are present, the regression coefficients are
usually extremely sensitive to k for small values of k. However, as k
increases, the coefficients stabilize. Hoerl and Kennard (1970a, b) recom-
mend selecting a value of k in the stable region within which moderate
changes in k do not affect the coefficients significantly.
A diversity of automatic statistical approaches for selecting k have
been discussed in the literature, e.g., by Dempster et al. (1977), and
Wichern and Churchill (1978). In the program documented herein, the user
has the option of displaying the solutions for a set of values of k or
Letting the program select k. The automatic method is based on both statis-
tical and physical criteria. The motivation for choosing the particular
criteria used is as follows. Typically, when serious multicollinearities
;u'. pi;?-• dV: , In the weighted least squares solution, one S. value in a
collinear set will be excessively large and another will be negative with
large magnitude. That is, one S. value will blow up on the positive side
J
and another will blow up on the negative side. Simply setting the negative
values equal to zero, therefore, would solve only half of the problem.
As k increases, the negative values usually approach zero, and some or
all of them may become positive. Thus, one criterion is to select k so
as to minimize the magnitudes of the negative S. values.
Another criterion is to optimize the agreement between the observed and
calculated aerosol concentration for the size fraction being considered.
This criterion involves the complete set of S. values and has a reasonable
physical basis. To address both criteria k is selected so as to minimize
the following:
s\
(C - CT)2 +UZ S.2 (9)
S <0 J
223
-------
where W IH a user-specified weighting factor. Increasing W increases the
relative Importance of avoiding negative, values with significant magnitudes.
Decreasing W increases the relative importance of achieving agreement
between the measured and calculated aerosol mass concentrations. The sum of
squares of the negative S. values is zero if there are no negatives.
While this is obviously only one of many possible schemes for selecting
k, it is based on reasonable physical and statistical criteria. Neverthe-
less, neither this nor any automatic scheme can guarantee selection of the
optimal value of k. However, the suggested scheme tends to select values
of k in the stable region. Thus, slight deviation from the optimal k should
not introduce significant errors in the solution.
Weighted regression, effective-variance calculations, and ridge re-
gression all offer enhancements to the ordinary multiple regression approach
to the CMB problem. Fortunately, these three features can be used simul-
taneously or in various combinations. In weighted regression, a transfor-
mation is first made to stabilize the error variance of the dependent
variable (Draper and Smith, 1966). This transformation can be made using
the original error variances of the dependent variable or the effective
variances. Subsequently, the regression analysis is performed using the
transformed variables; this can be accomplished using either ordinary
regression or ridge regression. Finally, the necessary inverse transfor-
mation is performed to obtain the solution in the original data units.
The simultaneous use of the three features mentioned above requires cer-
tain minor changes to the procedures. For example, weighted ridge regression
with an intercept is most easily carried out if x'x and x'y are normalized
but not centered. Thus, x'x and x'y do not contain correlation coefficients,
as in conventional ridge regression practices discussed above. However, the
diagonal elements of x'x are unity, as in the conventional case. Thus, as
before, a k value of 0.1 represents a 10 percent perturbation of the diagonal
elements. This and other minor procedural changes do not impact the inter-
pretability of the results.
224
-------
In the discussion above, the expression for the covariance matrix of
the standardized regression coefficients is given in Eq. 8. The quantity
02 is the error variance of the dependent variable; a common error variance
applies after the transformation to stabilize the variance has been
performed.
The quantity 02 is estimated by s, the conventional standard error:
s2
n+1 ^
I (Y± - Y±)2
izi - (10)
n+1 - L
where
Y. is the i value of the dependent variable, i.e., Y. is the measured
I" Vi
concentration of the i species, 1
-------
TECHNICAL REPORT DATA
(flease read Instructions on the reverse before completing)
. REPORT NO.
EPA-450/4-83-014
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Receptor Model Technical Series
Volume III
User's Manual for Chemical Mass Balance Model
5 REPORT DATE
July 1983
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
Hugh J. Williamson and Dennis A. DuBose
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Radian Corporation
PO Box 9948
Austin, Texas 78766
11. CONTRACT/GRANT NO.
68-02-3513
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Research Triangle Park, North Carolina 27711
13. TYPE. OF REPORT AND PERIOD COVERED
Final
A. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: Warren P. Freas
16. ABSTRACT
In recent years there has been increasing interest in source apportionment of
ambient aerosol concentrations through chemical mass balance (CMB) analysis. This
report discusses CMB analysis through weighted least squares with options to include
effective variance and ridge regression features. Effective variances are refined
estimates of the weights employed in weighted least squares.
This report documents an interactive FORTRAN computer program which performs
aerosol source apportionment through the analysis methods discussed above. The
original version of the program, which performed weighted least squares with the
effective variance option, was developed at the Oregon Graduate Center based on the
Doctoral Dissertation of Dr. John Watson. In the current version, a ridge regression
feature was added, along with various modifications intended to enhance the ease of
use of the program.
17.
a.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Receptor Model
Chemical Mass Balance
Source Apportionment
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
233
20. SECURITY CLASS (Thispagcl
Unclassified
22. PRICE
EPA Form 2220-1 (R»v. 4-77) PREVIOUS EDITION is OBSOLETE
------- |