OCR error (C:\Conversion\JobRoot\00000AXA\tiff\2000XFXN.tif): Unspecified error
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are.
1 Environmental Health Effects Research
2 Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects, assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-118
July 1978
INTERCOMPARISON OF SAMPLERS USED IN THE
DETERMINATION OF AEROSOL COMPOSITION
David C. Camp and Allan L. Van Lehn
Lawrence Livermore Laboratory
Livermore, CA 94550
and
Billy W. Loo
Lawrence Berkeley Laboratory
Berkeley, CA 94720
EPA-IAG-D7-F1108
Project Officer
Thomas G. Dzubay and Robert K. Stevens
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, NC 27711
This study was conducted
in cooperation with the
U.S. Department of Energy
Washington, D.C.
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH & DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation. 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 re-
commendation for use.
-------
ABSTRACT
An intercomparison study was carried out to evaluate the performance of
11 different designs of aerosol samplers. The samplers were operated by
participating scientific groups having recognized expertise in sampler de-
velopment, operation, and the subsequent analysis of the samples they col-
lected. The devices tested include hi-vol, TWO MASS, multiday, cyclone,
CHAMP, streaker, stacked filter, and manual and automated dichotomous
samplers. The samplers were operated simultaneously (on the roof of the
Federal Building) in Charleston, West Virginia for eight consecutive days
during May of 1977. The collection surfaces of each sampler were changed
at least every 12 hours which enabled the intercomparison to be made for
16 sampling periods. Samples collected were analyzed by each participant
for one or more of the pollutants for mass, nitrate, sulfur or sulfate,
lead, and 8 other trace elements. The analysis methods used by the various
participants included x-ray fluorescence, particle induced x-ray emission,
atomic absorption, ion chromatography, colorimetry, beta gauge and gravi-
metric techniques.
Most of the samplers separated the aerosol into two fractions with 50
percent separation diameters ranging from 2.4 pro to 4.3 jjm. The upper 50
percent cutoff diameter for the various samplers ranged from 14 jjm to above
30 pm. Data were intercompared for the total, fine and coarse size frac-
tions.
Best agreement among samplers was found for elements such as sulfur and
lead that occurred primarily in the fine fraction.
The amount of total mass collected appeared to be strongly influenced
by the upper 50 percent cutoff diameter of each sampler.
For the stacked filter samplers and the tandem filter samplers, the
fine fraction appeared to be enriched with crustal elements such as Si, Ca,
and Fe, which suggests that there are particle bounce errors.
Of all the samplers reporting results, the automated dichotomous
sampler showed the greatest precision for the fine fraction species.
In the intercomparison study, total sulfur was measured by x-ray
fluorescence, and sulfate was measured by ion chromatography and by wet
chemical methods. The mean sulfur and sulfate concentrations were in agree-
ment with the assumption that all of the sulfur is in the form of sulfate.
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgements viii
1. Introduction 1
2. Conclusions 9
3. Recommendations 11
4. Field Study Conditions 13
5. Mathematical Treatment of Results 19
6. Selected Results and Discussion 27
References 49
Appendices
A. Sampler and Analytical Technique Summaries 51
B. Outlier and Regression Analysis Equations 94
C. Tabulated Pollutant Results 99
-------
FIGURES
Number Page
1 Photograph of the samplers on the roof of the Federal Building,
and a sampler identification chart 6
2 Hourly temperatures recorded at the Charleston Airport during
the study 15
3 Hourly relative humidity recorded at the Charleston Airport
during the study 15
4 The mean hourly wind direction and velocity recorded atop the
Federal Building 17
5 Wind speed and direction at the study shown as wind roses .... 18
6 A portion of one of the tabular results listed in Appendix C . . . 25
7 Mean mass and selected pollutant concentrations vs period .... 31
8 Average sampler ratios vs sampler type for selected pollutants . . 37
9 Average sampler ratios for sulfur and sulfate combined 40
10 Average sampler ratios for zinc and copper 45
VI
-------
TABLES
Number Page
1 Intercomparison Study Participants ....... 3
2 Sampler Types Brought to the Study 5
3 Sampler Types, Inlet Heights and Spacings ... 7
4 Sampler Types vs. Operating Characteristics and Pollutants
Reported 8
5 Outliers According to the Dixon Criterion (a = 4%) 22
6 Mean Concentration Changes vs Number of Samplers Averaged .... 23
7 16-Period Pollutant Concentration Averages from all Samplers
Reporting Results for May 11-19, 1977 29
8 Possible South to East Dispersion in the Air Sampled 32
9 Regression Analysis Fits of Referee Analysts Results 33
10 Ratio of XRF Results for Cut to Uncut Filter Samples 34
11 Participant Results from Analytical Technique Quality Control
Samples 35
12 Pairwise Intercomparison of Sampler Precisions (%) 39
13 Coarse Particle Bounce into the Fine Fraction 43
14 Percent of Pollutant in the Small or Large Fraction 48
-------
ACKNOWLEDGMENTS
It is a pleasure to acknowledge the essential guidance provided by Mr.
William Matthews, West Virginia Federal Buildings Manager, during the planning
phase of this study. He was instrumental in seeing that additional electrical
power was installed on the roof of the Federal Building in Charleston, WV.
The cooperation and assistance rendered by Mr. Howard Stanley, Charleston
Federal Building Foreman, and his crew allowed the study to proceed smoothly.
The continued support, creative suggestions, and encouragement offered by
the Project Officers, Mr. Robert Stevens and Dr. Tom Dzubay, in the planning
phase, during execution, and especially during the preparation of the final
report is gratefully acknowledged. They were also helpful in all phases of
the November, 1977 Workshop at which preliminary results of this study were
presented.
Grateful appreciation is also extended to each of the intercomparison
study participants who, through their creative criticisms, helped make this
final report more comprehensive and complete.
Finally, it is a pleasure to acknowledge the Staff of the Director's
Office Correspondence Center at Lawrence Livermore Laboratory who so cheerful-
ly typed the many drafts and revisions required by a report of this magnitude.
vm
-------
SECTION 1
INTRODUCTION
HISTORICAL PERSPECTIVE
One of the consequences that grew out of the environmental concerns
of the 1960's was a demand for better air quality. For research scientists
in this field this meant improving the measurement, characterization, and
understanding of air pollutants both at their emission sources and in the
ambient atmosphere. Consequently, more attention began to be focused on
the performance of aerosol particulate samplers. Performance parameters
examined more closely included sampler inlet designs, their flow rate capa-
bilities, the particle size cutpoints employed, internal particle loss mecha-
nisms, and types of collection substrates used, to name just a few. As a
result of this attention, the development of both simple and sophisticated
aerosol particulate samplers began to take place all across the country.
Several conferences (1,2) helped focus on what had been accomplished as well
as on what the needs were. By 1976 a great deal of air pollution information
was being collected with the many different sampler designs, and results
were being published in numerous journals. One of the questions that arose
following the second conference (2) was how intercomparable are the results
from different samplers? If the results did differ, was it by small percent-
ages, by factors of 2, or did they differ by orders of magnitude? Were some
pollutants easier to measure than others? Thus, it was suggested that a
study be carried out which would begin to answer these and other questions.
OBJECTIVE
The objective of the study was to operate representative state-of-the-art
aerosol particulate samplers simultaneously and side-by-side for an extended
period and to compare the analytical results. This objective was achieved
through a field study at which the samplers were operated by their developers
or users in nearly the same manner as they are operated in typical air pollu-
tion research and/or monitoring programs. The samples collected were analyzed
for selected pollutant concentrations by each participant using those analyti-
cal methods normally employed for the type of sample and pollutant collected.
SITE SELECTION
Several considerations entered into the selection of an appropriate site
for the field study. Among the more important were that sufficient space,
sufficient electrical power, and security for the samplers be provided. It
was also desirable to conduct the study in an area where some pollutant infor-
mation was available. The earlier results would give some indication of the
1
-------
kind of pollutant concentrations to be expected. These considerations led to
the selection of the Federal Building in downtown Charleston, West Virginia,
as the site where the field study would be conducted.
This is a five story building through which there is access to a large,
open roof area. It has the added bonus of a 24-hour security service. The
earlier study (3) carried out in the Charleston area indicated that the month
of May offered satisfactory weather conditions for the study. The field study
was scheduled to begin at 8 am, Wednesday, May 11, 1977, and end at 8 am,
Thursday, May 19, 1977. Sample collection was to be continuous throughout
the study with collection substrates to be changed at least once every 12
hours. This 8-day duration led to sixteen 12-hour periods for which pollutant
concentrations could be intercompared.
POLLUTANTS SPECIES SELECTED
There are a wide range of air pollutant species that could have been
measured in such a study. Because the objective of the study was to intercom-
pare results from as many different types of samplers as possible, the number
of pollutant species selected was purposefully restricted.
Those pollutant results for which an intercomparison was believed to be
meaningful were mass, nitrate, sulfur and sulfate, lead, and eight other
trace elements to be selected after the field study was completed. Since the
type of samplers brought to the study varied from simple to complex, the study
was designed so that comparisons could be made for the total pollutant concen-
tration reported; and also for fine and coarse fractions for samplers which
collected particles according to aerodynamic size.
Following completion of the study eight trace elements were chosen.
Three of these - silicon, calcium, and titanium - are representative of those
elements found in the earth's crust. A fourth element, iron, is also found in
this group, but may also be associated with local emission sources. These
four elements tend to concentrate in the coarse or large particulate fraction,
that is, in particles having aerodynamic sizes greater than 2.0 micrometers.
Three other trace elements chosen were copper, zinc, and bromine. The last
was chosen because of the well known bromine to lead ratio characteristic of
automotive exhausts. The eighth trace element chosen was selenium, which usu-
ally occurs in very low concentrations. Its measurement becomes a challenging
test of the various samplers and analytical methods used in the study.
PARTICIPANT SELECTION
Because of space restrictions atop the Federal Building in Charleston,
only a limited number of samplers could be accommodated. This restriction
plus the desirable option of replicate samplers limited the number of partic-
ipants that could be invited to the study. Table 1 lists those participants
invited to the field study and who brought samplers which were representative
of those aerosol collection devices most widely used in research and monitor-
ing programs.
-------
TABLE 1. INTERCOMPARISON STUDY PARTICIPANTS
Mr. Robert Burton:
Dr. Thomas Cahill
Dr. R. Delumyea,
and Dr. E. Macias:
Dr. Thomas Dzubay
and R. K. Stevens:
Dr. Martin Hudson:
Dr. Billy W. Loo:
Dr. Peter Mueller:
Dr. Charles Rodes:
Dr. Roger Tanner and
Dr. Leonard Newman:
Health Effects Research Laboratory,
Environmental Protection Agency,
Research Triangle Park, NC 27711
Physics Department, University of California,
Davis, CA 95616
Department of Chemistry, Washington University,
St. Louis, MO 63130
Environmental Sciences Research Laboratory,
Environmental Protection Agency,
Research Triangle Park, NC 27711
Physics Department, Florida State University,
Tallahassee, FL 32306
Lawrence Berkeley Laboratory, Berkeley, CA 94720
Environmental Research and Technology, Inc.,
West Lake Village, CA 91361
Environmental Monitoring & Support Laboratory,
Environmental Protection Agency,
Research Triangle Park, NC 27711
Atmospheric Sciences Division,
Brookhaven National Laboratory,
Upton, L.I., NY 11973
SAMPLER TYPES AND ARRANGEMENT
Table 2 lists those samplers brought to the study. Both a one page sum-
mary and detailed description of each type can be found in Appendix A. It
is apparent that a number of duplicate samplers were brought. The well known
high volume sampler was well represented at the study. Included were those
with and without particle size inlet restrictions as used in the CHESS* study
carried out in a number of communities by the Environmental Protection Agency
(EPA). Of the 35 samplers brought to the study, 20 had the capability of
collecting particles in two or more size fractions. Of these 15 reported size
fraction results.
All of the samplers assembled on the roof of the Federal Building are
shown in Figure 1. This view looks southeast toward downtown Charleston, WV.
* Community Health and Environmental Surveillance Studies
3
-------
The lower part of Fig. 1 includes a sketch of each sampler type. A number is
is used to identify each sampler or group of samplers, while a capital letter
is attached to only some of the samplers. This letter identification denotes
those samplers reporting results which were intercompared and is useful in
distinguishing between duplicate or triplicate samplers brought by the same
participant. Sampler A is located approximately 9 meters from the eastern
corner of the building while sample U is about 9 meters fro.n the southern cor-
ner of the Federal Building. All of the samplers were located on a line
approximately 1.5 meters from the southeast face of the building. All around
the edge of the roof the side of the building continues upward 0.4 meter above
the roof level.
Table 3 lists all of the samplers or groups assembled on the roof ac-
cording to the numerical order shown on the sketch in Fig. 1. The last two
columns list the sampler inlet height to the nearest centimeter, and the ap-
proximate spacing between sampler inlets. Inlet heights varied from a low of
about 0.9 meters to a high of 1.75 meters. (This excludes the S02 sampler in-
lets at 2.4 m. Its results were not part of the intercomparison study).
Table 4 lists the more important operating characteristics for those
samplers whose reported results were intercompared. The alphabetical identi-
fication of the various samplers is used throughout this report both in the
graphical presentations and in the tabular intercomparisons. There are sever-
al important points to be noted in this table. First, an examination of the
column listing the fine cut points reveals that they vary between 2.5 and 4.3
micrometers. The first two stages of Sampler G, the Sierra Multiday Sampler,
were added together before comparing its results with other samplers. Simi-
larly, the column listing the inlet cutoff also shows a range of large parti-
cle sizes accepted by each sampler. These variations in cut-point sizes must
be kept in mind when examining the results of the fine and coarse particulate
fractions. The column listing the frequency of filter changes shows that for
some samplers, two to four sub-period totals had to be added together to ob-
tain the desired 12-hour pollutant concentrations. Florida State University
employs an analytical technique which is capable of analyzing their linear
streaker filters to within + 30 minutes.
The filter substrates used in the automated dichotomous samplers - C, L,
and S - were changed automatically every 6 hours. These samplers also employ-
ed a flow control mechanism which shut the samplers down in the event the flow
rate dropped below a prescribed tolerance. On several occasions at night the
particulate loading was sufficiently heavy to cause shut down very near the
end of the first 6 hour period. In these instances the following 6 hours of
data were not obtained.
-------
TABLE 2. SAMPLER TYPES BROUGHT TO THE STUDY
a)
Type of Sampler Participant Inst.
Diffusion Battery Sampler
High Volume Sampler:
High Volume Samplers:
High Volume Sampler:
Champ (with size cuts)
Dichotomous Samplers - ERC.
Dichotomous Samplers - Manual
Dichotomous Samplers - Automated
Linear Streaker
Batelle Impactors
Tandem Filter Units
Stacked Filter Units
Sierra Multiday
Cyclone (1 w & 1 w/o size cut)
Two Mass
Sulfur dioxide experiment
Tanner
Tanner
Rodes
Burton
Burton
Rodes
Dzubay
Loo
Hudson
Hudson
Dzubay
Cahill
Cahill
Mueller
Del umyea
Loo
BNL
BNL
EPA
EPA
EPA
EPA
EPA
LBL
FSU
FSU
EPA
UCD
UCD
ERT
WU
LBL
# Units
at study
1
1
2
1
2
2
2
4
4
4
2
2
1
2
4
4
Results
reported for
(b)
1
lc)
1
2
lc)
2
3d)
2e)
(e)
2
2f)
1
2
1
(g)
a) See Appendix A for a detailed description of each sampler type.
b) Results were reported but were not intercompared, and are not included
in the tables of Appendix C.
c) Data were taken with both samplers for only half of the study.
d) The fourth sampler was included in case spare parts were needed for the
other three.
e) All four operated throughout the study. Results available from FSU.
f) A third total filter unit was operated throughout the study with no
sizing. Only one filter change was made after four days, hence its
results could not be intercompared.
g) These samplers were not part of the intercomparision study.
-------
CD
O)
QJ ID
-------
TABLE 3. SAMPLER TYPES, INLET HEIGHTS AND SPACINGS
Sampler Type
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.
'Hi-Vol and Diffusion
Champ + Size Fractionator
Auto Dichotomous
Manual Dichotomous
Cyclone-total
Cyclone-fine
Sierra Multiday
Total Filter Unit-1 stage
Stacked Filter Unit-2 stages
Stacked Filter Unit-3 stages
4-Batelle Impactors 6 stages
4-Linear Streakers
Sulfur Dioxide Analyzer
ERC Sampler, w/o MgO
ERC Sampler, w MgO
ERC Sampler w/o MgO
Auto Dichotomous
2-Two Mass Samplers
Stacked Fil ter Unit
Stacked Filter Unit
Hi-Vol
ERC Virtual Impactor
ERC Virtual Impactor
Hi -Vol
Two Mass + Beta Gauge
Manual Dichotomous
Auto Dichotomous
Auto Dichotomous
Hi-Vol
Champ + Size Fractionator
Investigator Inlet Height(cm) Spacing(cm)
(Tanner-BNL) 135
(Burton-EPA) 160
(Loo-LBL) 160
(Dzubay-EPA) 132
(Muller-ERT) 175
(Muller-ERT) 175
(Cahill-UCD) 140
(Cahill-UCD) 137
(Cahill-UCD) 140
(Cahill-UCD) 140
(Hudson-FSU) 165
(Hudson-FSU) 185
(Loo-LBL) 114
(Loo-LBL) 238
(Loo-LBL) 238
(Loo-LBL) 107
(Loo-LBL) 160
(Delumyea-WU) 89
(Dzubay-EPA) 157
(Dzubay-EPA) 157
(Rodes-EPA) 112
(Rodes-EPA) 124
(Rodes-EPA) 124
(Rodes-EPA) 112
(Delumyea-WU) 114
(Dzubay-EPA) 132
(not working) 157
(Loo-LBL) 157
(Burton-EPA) 112
(Burton-EPA) 157
& 114 86
91
69
81
71
71
20
30
28
•v. 81
140
^ 0
36
36
76
71
84
56
124
74
56
89
132
91
71
58
79
64
a'These numbers are used in Figure 1 to identify the sampler groups.
-------
TABLE 4. SAMPLER TYPE VS. OPERATING CHARACTERISTICS & POLLUTANTS REPORTED
Sampler Type Investigator Fine Cut3'
and Lab Points
urn
Inlet3*
Cutoff
urn
Flowa) Filter3*
Rate Change
(l/m) (hrs)
Pollutants Reported
M N S sul L trace"
a 0 0 fur e ele-
s 3 4 a ments
s d
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
Hi-Vol
Champ
+ Size Cut
Auto Dicho-
tomous
Manual
Dichotomous
ERT - Total
ERT - Fine
Sierra
Multiday
Total Filter
1 stage
Stacked Filter
2 stages
Linear Streaker
Linear Streaker
Auto Dicho-
tomous
Tandem Filter
Tanden Filter
ERC Dichotomous
Hi-Vol
Two-Mass
Manual
Dichotomous
Auto Dicho-
tomous
Hi-Vol
Champ
+ Size Cut
Tanner
Burton
Loo
Dzubay
Mueller
Mueller
Cahill
Cahill
Cahill
Hudson
Hudson
Loo
Dzubay
Dzubay
Rodes
Rodes
BNL
EPA
LBL
EPA
ERT
ERT
UCD
UCD
UCD
FSU
FSU
LBL
EPA
EPA
EPA
EPA
Delumyea WU
Dzubay
Loo
Burton
Burton
EPA
EPA
EPA
EPA
3.5
2.4
3.5
-
2.5
0.78
4.3
-
2.6
-
-
2.4
3.5
3.5
3.5
-
3.5
3.5
2.4
-
3.5
large
-x, 26
•v, 25
14
'vlB
-
<20
<20
<20
-v-15
-v-15
^25
c)
c)
•v20
large
c)
14
-v-25
large
•\. 26
700
1130
50
14
100
100
24
24
5
0.35
0.50
50
7.2
5.9
14
1415
16.7
14
50
1130
1130
12
12
6
12
3
3
12
12
12
N S
M N S
M
M S
M N S
M N S
continuous
continuous
6
12
12
12
12
3
12
6
12
12
M
M S
M S
M
M N S
M
M N S
M N S
s
s
s
s
s
s
s
s
s
s
s
s
s
L
L
L
L
L
L
L
L
L
L
L
L
L
te
te
te
te
te
te
te
te
te
te
te
te
a) See Appendix A for a more detailed discussion of these sampler types and their
operating characteristics.
b) Trace element means that silicon, calcium, titanium, iron, copper, zinc, bromine,
and selenium were usually reported when they could be detected.
c) Not measured
-------
SECTION 2
CONCLUSIONS
An intercomparison study was carried out to evaluate the performance of
11 different designs of aerosol samplers. The devices tested include hi-vol,
TWO MASS, cyclone, CHAMP, streaker, stacked filter, and manual and automated
dichotomous samplers. The samplers were operated simultaneously on the roof
of the Federal Building in Charleston, West Virginia for eight consecutive
days during May of 1977. The collection surfaces of each sampler were changed
at least every 12 hours which enabled the intercomparison to be made for 16
sampling periods. The collected samples were returned to the laboratory of
each participant and analyzed for mass, nitrate, sulfur or sulfate, lead, and
8 other elements.
Most of the samplers separated the aerosol into two fractions with 50
percent separation diameters ranging from 2.4 pm to 4.3 jum. The upper 50
percent cutoff diameter for the various samplers ranged from 14 urn to above
30 urn. Results were intercompared for the total, fine and coarse size
fractions. Elements found most abundantly in the fine fraction included S,
Zn, Cu, Br, and Pb; elements found most abundantly in the coarse fraction
included Si, Ca, Ti , and Fe. The total mass was equally distributed between
both size fractions.
Best agreements among samplers was found for elements that occurred
primarily in the fine fraction. For lead the standard deviation of the
results from all of the samplers were 17% and 11% for the total and fine
fractions respectively. The standard deviations for sulfur were 15% and 11%
for the total and fine fractions respectively. In computing the standard de-
viation for sulfur, the results associated with the TWO MASS sampler were
excluded. For that sampler the results were lower than the mean by a factor
of about 3.
The amount of total mass collected appeared to be strongly influenced by
the upper 50 percent cutoff diameter of each sampler. The amounts of Si, Ca,
Ti , and Fe collected in the coarse fraction were similarly affected by the
inlet. The hi-vol sampler, which had the largest upper cutoff diameter,
collected about 50% more mass than the manual dichotomous sampler, which had
the lowest upper diameter (14
For the stacked filter samplers and the tandem filter samplers, the fine
fraction was enriched in crustal elements such as Si, Ca, and Fe. This
indicates that there are particle bounce errors for such samplers.
Of all the samplers reporting results the automatic dichotomous sampler
9
-------
showed the greatest precision for the fine fraction species. Three automated
dichotomous samplers when compared pairwise yielded an average agreement of
1.8%, 1.8% and 3.3% for sulfur, lead, and mass in the fine fraction, re-
spectively. For the total fraction, the agreement between pairs averaged 2%,
3%, 7% and 12% for sulfur, lead, mass and iron, respectively.
In the intercomparison study, total sulfur was measured by x-ray
fluorescence, and sulfate was measured by ion chromatography and by wet
chemical methods. The mean sulfur and sulfate concentrations were in agree-
ment with the assumption that all of the sulfur is in the form of sulfate.
The nitrate results showed the greatest differences among samplers, and no
simple explanation for the differences was found.
10
-------
SECTION 3
RECOMMENDATIONS
This sampler intercomparison study was the first side-by-side field test
of a large number of representative state-of-the-art aerosol samplers. It was
successful in that it demonstrated for those samplers brought to the study
that particulate mass and selected pollutant concentrations can be determined
with reasonable accuracy. In fact, for lead, sulfur, and zinc concentrations
in the fine fraction, all reported results were within 10% of a mean value.
On the other hand nitrate concentrations were not well determined in this
study for reasons that remain unclear at this time.
Among those who participated in this study, it was unanimous that the
field study was well worthwhile, and it was a general consensous that a simi-
lar study should be held again at some future date. With such a future study
as a possibility and based on the experiences gained in the present study a
number of recommendations can be advanced which may help insure that any fu-
ture study will be even more successful.
The flow rate of all samples at any future study should be checked
periodically (e.g., at least twice a day). More importantly all flow rates
ought to be cross-checked with a single flow rate meter so that all sampler
flow rates at the study are tied relatively to a single calibration device.
All samplers brought to a future study should have nearly the same inlet
restrictions and preferably axially symmetric inlets. Similarly, the cut
point used to define the fine particle fraction for all samplers at such a
study should be very nearly the same.
Samplers used to evaluate the possibility of any inhomogeneity in the
sampled air should be run side-by-side both prior to and following the study
for periods approximately 1/4 the duration of the study (i.e., two days before
and two days after in an 8-day study).
One or more separate studies should be held soon and devoted to a de-
termination of those factors that influence the determination of nitrate
concentrations.
Following completion of the study, analytical technique quality control
samples should be mailed to a'l participants and all participants required
to report concentrations in ng/cm for at least three selected pollutants.
The method of appealing to independent referee analysts for quality control
was successful in this study.
11
-------
All samplers deployed at any future study should guarantee that they do
not exhaust any pollutant which can bias another samplers results. This may
require additional filtration of the exhausts of some samplers.
The spacing between samplers at any future study should average 1 meter.
All participants should be required to field and report results for at
least duplicate samplers. Also, no samplers should be allowed to take up
space or electrical power which do not report results.
As was the case in this study, some form of security is desirable to
prevent access to the samplers by the general public.
Multi-redundant fusing of the electrical power was well worth the plan-
ning and additional investment required. A number of samplers at this study
blew fuses at various times, but none was responsible for the shut down of
any other sampler.
Although no rain fell during this study, any future study should plan
for this eventuality.
12
-------
SECTION 4
FIELD STUDY CONDITIONS
AIR HOMOGENITY SAMPLERS AND QUALITY CONTROL SAMPLES
The aerosol samplers were assembled on the roof of the five story Federal
Building in Charleston, WV along the southeast face of the building. Since
the objective of the field study was to intercompare the pollutant concentra-
tions obtained by samplers that were not located at the same point, it was ne-
cessary to test the possibility that the air sampled might not be homogeneous.
Interaction between wind and the roof penthouse structures might create non-
uniform wind streams. To monitor this possibility, three Lawrence Berkeley
Laboratory (LBL) designed automated dichotomous samplers were operated near
both ends and near the middle of the linear array of samplers. A detailed
discussion of the results is presented in Section 6.
In general duplicate samplers brought by a single participant were not
set up side by side. However, an attempt was made to "pair-up" samplers which
were similar in operating or design principle. Two samplers might be operated
side by side, but were brought to the study by different participants. Some
types of samplers were more effectively operated in clusters.
In planning the study it was recognized that the concentrations reported
by the various participants would be different. How much of any reported dif-
ferences are due to sampler performances and how much to the different analy-
tical methods used to analyze the samples, should be resolved if possible.
The method chosen to resolve these two for this study was based, in part, on
the fact that most of the participants would report trace element concentra-
tions determined from an x-ray emission technique. The method chosen was as
follows.
From one of the automated dichotomous samplers operated at the field
study, a set of fine particulate fraction samples were selected. This set
was analyzed by two referee analysts who were not participants in the study.
Both referee analysts used energy dispersive x-ray fluoresence analysis to
determine the trace element concentrations in ng/cm of filter area. They
sent their results to the study director and from the combined data, concen-
tration values in ng/cm were adopted for each trace element and sample in
the set. Then, 3 samples were selected from the set and mailed to each of 8
participants (24 samples-total). X-ray fluorescence analysis carried out
before and after extremely vigorous shake tests on the entire set of samples
demonstrated that no particle losses would occur as a result of mailing. Each
participant then determined selected trace element concentrations in ng/cm2
for his three samples and reported these results to the study director. Their
13
-------
results were compared with the adopted values from the referee analysts.
Agreement between the two would suggest that the analytical method was not a
source of any reported pollutant concentration differences. Results for each
participants analytical quality control samples can be found in Section 6.
FIELD STUDY PRECAUTIONS AND PROCEDURES
The roof of the Federal Building was of a standard tar and gravel con-
struction. Since the filter substrates used in most of the samplers required
changing every 12 hours, a great deal of traffic would occur on this roof sur-
face. To reduce the potential for creating large particles, certain precau-
tions and procedures were adopted. A plywood walkway was laid down on the
roof all along and in front (NW) of the samplers. A narrow walkway leading
from the roof entrance/exit to the samplers was also laid down. Almost all
participant traffic on the roof used these walkways. The roof also contained
3 rest rooms vent stacks and 2 water drains in the vicinity of the samplers.
The presence of ^S gas was clearly noted from these stacks (most serious
in the former), hence they were vented away from the immediate sampling area
using 6.1 m lengths of flexible plastic pipe. The 3 rest room stacks were
vented at the top of the 1st level penthouse (3.35 m) and the water drains
were vented at roof level 4 m NW of the sampler array.
Filter substrates were changed at least every 12 hours, at 8 am and 8 pm.
Each participant was responsible for the operation of his own sampler, includ-
ing the changing and storage of new or used filter substrates. At other times
traffic on the roof was discouraged. Except for one TV news interview, there
was no public traffic on the roof throughout the study. At the end of the
study each participant was responsible for the disassembly and removal of his
samplers.
Electrical power for all of the samplers was distributed via 12 circuits
each with an individual 20 amp circuit breaker. Each of these circuits ter-
minated in a 4-plex electrical outlet into which were plugged one or two in-
dividually fused (at 15 amps) 6-circuit multi-outlet modules. Individual
sampler power cords were plugged into one of the six circuits which were also
separately fused at an appropriate current rating for the particular sampler
connected. In the event one sampler drew too much current, this multi-redun-
dant fusing isolated and protected each sampler from all others.
METEOROLOGICAL CONDITIONS
Meteorological conditions were satisfactory during the entire study;
however, no rain was recorded during any of the sixteen 12-hour periods. Fig.
2 shows the hourly temperature variations recorded at the Charleston Airport
located approximately two miles northeast of the Federal Building. Morning
temperatures varied from a low of 3°C (37°F) on the first day of the study to
17°C (63°F) on the last day of the study, while the daytime maximum readings
varied from 20°C (68°F) to 31°C (87°F). Since these data were obtained above
the valley floor, actual temperatures atop the Federal Building may have been
several degrees higher.
The relatively humidity as recorded hourly at the Charleston Airport
14
-------
90°
80°
70°
60°
50°
40°
30°
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
:.*
May 11-19, 1977
i i I i i l i i l i i l i i l i i l i i I i i l i i l i i l i i l i i l i i I i i l
SAM SAM SAM SAM SAM SAM SAM SAM SAM
8PM 8PM 8PM 8PM 8PM 8PM 8PM 8PM
Figure 2. Hourly temperatures recorded at the U.S. Weather Station, Charleston
Metropolitan Airport, throughout the intercomparison study.
100
90
80
70
60
50
40
30
20
10
n
123456789
i i | i i | i i | i i | i i | 1 1 | 1 1 | 1 1 A I !
*- ^
1 ui •
• **
**A & **
/ > m*^ .
.-A A
* •'* .-* :\ • -
^' A . A .' . A
""*•*•• «* * A
. .* • A
~~* « • * * » A
r\A ^A -J1 ^
^^ /. ^
_ ^
May 11-19, 1977
1 t I I I I I I I 1 1 I 1 1 1 | 1 1 1 1 1 1 1 1 1 1
10 11 12 13 14 15 16
I i i I i i I i i I i i I i i I i i I i i
• A A^
"^ i * .1
."A ; /A -
• A ' '
• —
* • * • ••
• • • . —
* ^\
* • • ^\ 7"^ T£L* —
A A ^t
zoyw
A A/
* •/
" Charleston airport •
Federal bldg A ~
i > * i i i l < < i i i i i , i , . i , ,
SAM SAM SAM SAM SAM SAM SAM SAM SAM
8PM 8PM 8PM 8PM 8PM 8PM 8PM 8PM
Figure 3. The relative humidity recorded hourly at the U.S. Weather Station,
Charleston Airport, and occasionally on the roof of the Federal
Building, downtown Charleston.
15
-------
varied from 18 to 9H, while that recorded intermittently atop the Federal
Building showed variations from 24 to 98%. Even though these ranges are dif-
ferent, it appears that the daily cycle of minimum humidity near 4 pm and
maximum near 8 am are sufficiently well represented by the hourly Airport
data. These data can be used to evaluate any humidity induced chemical ef-
fects that may have occurred in the filter substrates. The hourly readings
are shown in Fig. 3.
The wind speed and direction (as recorded by an anemometer, wind vane,
and strip chart recorders atop the Federal Building) showed a diurnal pattern.
The daytime wind direction averaged about 300° + 45°, while the nightime wind
direction was 120° + 15°. During four of the study periods (2,4,14,16) the
winds came nearly directly into the southeast face of the building. After the
study, the wind vane strip chart recording was examined and wind directions
were analyzed for every 15 minutes throughout the study. From this analysis,
hourly directional averages were obtained and plotted. The results are shown
in the top section of Fig. 4. Using the anemometer strip chart recording
the approximate mean wind velocity sustained during each hour was estimated
and also the maximum wind velocity recorded during each hour was noted. These
data are plotted in the lower portion of Fig. 4. Wind rose patterns are shown
in Fig. 5. The odd periods (daytime) are grouped together at the top, while
the even periods (night) are grouped at the bottom. These wind rose data
clearly show the diurnal wind pattern typical of the Kanawha Valley in the
spring.
16
-------
i i 11 i [ i i i i i [ r t i r ( | i i i i i | ii i i i | y
-o
OJ
CD
17
-------
r\
( 8 y
\ CN
CM
c/i
OJ
in
O
S-
c
•r—
2
t/i
re
E
O
_C
c/1
en
CD
CD
T3
CU
OJ
Q.
O
+J
re
-a
a;
•o
o
o
a;
01
Q.
to
a>
a>
en
18
-------
SECTION 5
MATHEMATICAL TREATMENT OF RESULTS
CALCULATION OF A MEAN CONCENTRATION
Table 4 listed 21 samplers that reported pollutant concentrations which
could be intercompared. Concentrations were reported in micrograms per cubic
meter for the mass fractions, and in nanograms per cubic meter for all other
pollutants. Since no sampler brought to the study was to be regarded as
the standard of comparison, the reported pollutant concentrations from all
samplers were assigned equal weights. Thus, a mean concentration can be cal-
culated for each of the 16 periods, for the 13 different pollutants, and for
the total, fine, and coarse particulate fractions. Only one participant re-
ported results for coarse nitrate, while thirteen participants reported total
lead concentrations.
Since all samplers were accorded equal weight, the calculation of a mean
concentration can be regarded as a "composite sampler" concentration. How-
ever, there are three effects which can bias the calculated mean concentra-
tions.
First, any sampler can potentially malfunction with the result that less
than 16 concentrations values are reported. The remaining values may or may
not distort the composite sampler concentration and a priori there is no way
to judge which samplers should be included or excluded. Another reason for
unreported results might be that the pollutant could not be detected because
of poor analytical sensitivity or insufficient amounts collected. Thus, not-
detected is not synonymous with zero concentration.
Second, in any large data set there always exists the possibility of out-
liers. These values may appear "bad" because of sampling errors, analytical
technique calibration errors, sampler malfunction, or errors in transcribing
results. Also an actual measured value may be quite reasonable, but is iden-
tified as an outlier by any statistical tests, because all other values re-
ported happen to be nearly identical. Thus, statistical tests for outlying
values should not be applied indiscriminately, but can be used to identify
unusual values in very large data sets.
Third, in calculating a composite concentration, the mean value may be
biased by duplicate or triplicate samplers if the reported concentrations from
these additional samplers all lie to one side of the mean. Thus, mean concen-
trations can be calculated from all samplers reporting results, or from only
one sampler of each type. The influence of these three factors (1) fewer than
16 periods of results reported, (2) outlying values, and (3) multi-sampler
19
-------
bias must be considered in calculating a mean. After defining the equations
from which the mean composite concentration is derived, each of the three
factors will be discussed.
The mean concentration for any period can be calculated from the simple
expression
*h h 4* h
where C^ is the concentration reported by the i sampler for the j period,
n is the number of sampler result sets used, and j is the period number (j =
1, 2,....16). The standard deviation associated with Z,- is calculated from
Szj = I rV I >, C,..4- - n
(2)
There will be 16 Z-mean concentrations for the 39 pollutant size fractions
reported (13 pollutants x 3 size fractions).
Fewer Than 8 Results
Because of a number of reasons including sampler malfunction, non-
detectable pollutant concentrations, and others, some participants reported
fewer than the total of 16 periods expected. Since there is no way to decide
a priori which results are "bad" or "good" or for what reasons, the decision
was arbitrarily made to eliminate from the mean concentration calculation any
result set containing fewer than 8 of the 16 period results for any given
pollutant size fraction. Applying this criterion, there were only 4 result
sets eliminated out of 139 reported for the total particulate fractions, 4
out of 111 for the small fraction, and 5 out of 100 for the large fraction.
Seven of these were because of low to non-detectable pollutant concentrations,
leaving only 6 of 350 (< 2%) result sets rejected because of fewer than 8
periods reported.
Outliers
Almost always in large data sets, values can be found which fall well
outside the expected statistical spread in the mean. For whatever reasons
these values appear "bad", two major decisions are required. First, should
a statistical test for outliers be applied and if so which one? Second, if
outliers are found, should they be eliminated from the mean indiscriminately,
or should some subjective judgement be employed? This is not simply rhetoric
since the definition of an outlying value depends on the breadth of the dis-
tribution deemed acceptable. If the probability for rejecting an observation
is made sufficiently small enough, all values can be made to be acceptable.
Two different statistical tests were used to identify outliers. One is
the well-known Dixon criterion (4); and the second, the Grubbs-criterion as
developed by Tietjen and Moore (5) for the detection of multiple outliers.
The latter is a multi-pass or cylical test in which means are calculated,
20
-------
outliers identified, the mean value recomputed, addditional outliers sought,
and etc. The number of cycles and the statistical breath to the mean distri-
bution depend on the number of observations. Application of this statistical
test resulted in many values being identified as "outliers" - too many it
seemed. Although it appears to be a very sophisticated test, the number of
observation sets (result sets) obtained in this study was too few for most of
the pollutant species reported. The Dixon criterion when used in the single
pass mode is much more conservative, i.e., fewer outliers identified. The
Dixon criterion was used with a arbitrarily set equal to 4%. This is the
probability assumed in rejecting an observation that really belongs in the
group. Any values falling more than 2% high or 2% low will be rejected.
The equations which define the Dixon criterion as applied to this study's
results can be found in Appendix B.
The Dixon criterion was applied to all result sets submitted. Immediate-
ly, punch card errors and recording errors were spotted. Several participants
submitted results for the wrong pollutant, or results in ng/cm instead of
ng/m . These were quickly identified by the outlier criteria. Another parti-
cipant had 9 of 16 results identified as outliers. Since only 7 of the 16
periods results remained, their results were eliminated from the mean compo-
site calculation. This was the only data set eliminated by the outlier test.
Table 5 lists all those sampler-period results which were identified as
outliers by the Dixon criterion. Most of these outlying values can be under-
stood or explained. For example, the selenium concentrations lie near the
threshold of detectability, hence wide flucuations are expected. The zinc
values occasionally clustered very tightly, allowing even a reasonable value
to appear as an outlier. The copper excursions were due to local sources
(copper from pump motors), and the excursions observed in certain crustal
elements were probably real. The outlying small nitrate values also could
be explained (see Section 6). In effect, then, except for the Q-sampler's
small sulfur results, which were generally low by a factor of 3, all of the
outliers identified were allowed to remain in the calculation of the mean.
Multiple Sampler Bias
There is a potential problem due to some participants having duplicate
samplers and others having only a single sampler. If one or more sets of
duplicate samplers lie well to one side of a calculated mean, the mean would
be biased. To examine this posssibility an alternative mean concentration
can be calculated by averaging results from only one sampler of each type.
Clearly, the latter mean is defined by a sub-set of the samplers defining the
former mean. Statistically, one expects the difference between these two
means to show a positive or negative bias if the additional duplicate samplers
are high or low. Since the number of samplers contributing to the the subset
mean is less than or equal to the full set, the standard deviation in the sub-
set mean is usually greater. Defining the full set of sampler means as Z and
the subset sampler means as X, the quantities (Z-X) and ($X-SZ) can be calcu-
lated and compared.
Table 6 shows the results of 592 mean comparisons. Almost as many mean
pollutant concentrations decreased as increased when the duplicate samplers
21
-------
were eliminated. As expected $x is on the average, larger than S . There-
fore, there is no strong evidence that including duplicate and triplicate
sampler results biases the composite Z-means. In general, samplers N and R
were lower than the composite mean by about the same amount that samplers
C and S were higher, hence these four samplers offset each other and reduced
the influence of other duplicate samplers.
TABLE 5. OUT1IERS ACCORDING TO DIXON CRITERION (a -
Total Sampler-Period
so4
S
Si
Fe
Cu
Zn
Se
Pb
a)
b)
c)
d)
e)
f)
P-13 a)
K-7 b>
J-9 d>
N-8
D-9 e)
N-8
M-2,9,11,14 f)
J-9; K-9 b)
Small Sampler-Period
N03 F-3,5,8,14
Large
Mass
S Q-4-7,9,12-15 c) Ti
Si 1-9,12 d)
Ti M-12;I-8,12
Zn N-8; S-3
Se M-9, 11,16; N-8
Fe
Cu
Zn
f) Se
Sampler-Period
0-8,10 a)
J-15 d)
N-8
D-2,9 e)
R-13
M-2 f)
EPA (Rodes) Hi-Vol Sampler. Reason for high value unknown.
FSU Linear Streaker. Apparently an artifact of the averaging process.
These data were low because of an unresolved analytical error.
Crustal elements. Incorrect subtraction of blank values.
Occasional high values of copper due to local sources, see Section 6.
Because selenium concentrations were at the threshold of detectability,
these values cannot be considered outliers.
RATIO TO MEANS
Once a composite mean has been defined for each pollutant specie and
sampling period, it is very useful to normalize all reported concentrations
to these composite sampler mean concentrations. This results in a table in
which all of the entries are near 1.0. This has numerous advantages. Among
them are
(1) Deviations from the mean are more easily observed and can be quickly
converted to approximate percentage differences.
(2) Average ratios for all 16-periods for each sampler can be calculated
independent of day to day changes in concentrations. A relative
measure of a sampler's overall accuracy (how far from 1.0 which is
equivalent to the mean concentration) is shown by this average; and
22
-------
an indication of a sampler's precision may be shown by the standard
deviation in the 16-period average ratio.
(3) Appropriate graphs of these average ratios vs samplers reporting
allow sampler to sampler comparisons.
(4) Multiple graphs of the type described in (3) allow pollutant to
pollutant comparisons independent of their actual aerosol con-
centration differences (which may be orders of magnitude).
These advantages will become evident upon a detailed examination of the re-
sults presented in Section 6.
TABLE 6. MEAN CONCENTRATION CHANGES VS NUMBER OF SAMPLERS AVERAGED
Quantity ____ Change _ Totals
a)
b)
(Z-X)
(sx - sz)
+ 0 a)
263 77 252 592
344 59 189 592
No change is defined as less than a 1% change in the Z-mean or standard
deviation, S.,, compared to the X-mean or Sx.
A total of 624 period-pollutant-means can be calculated (16 pds x 13
pollutants x 3 size fractions). Since only one sampler reported results
for large ^0^ and one sampler remained in the X-group for large sulfate,
only 592 comparisons were made.
REGRESSION ANALYSIS
Regression analyses were performed between the results from every pair
of samplers, and between each sampler and the composite Z-mean. If a linear
mathematical relationship is assumed to exist between two variables (sampler
concentrations), then we can write
y = a + bx , (3)
where y is the dependent variable, and x is the independent variable (the Z-
mean, for example). Since in this study both x and y are pollutant concentra-
tions in A*g/m or ng/m, the intercept, a, will also be in these units.
Standard equations exist for calculating ea, the error in a. Similarly, the
slope, b, and its error, 6k, can also be calculated. The equations for these
calculations are given in Appendix B. The extent to which artifacts exist
between any two pairs of samplers, will be indicated by the extent to which
a + ea does not overlap 0, and any consistent bias between any two samplers
by the extent to which b + e^ does not overlap 1.0. A measure of one samp-
lers precision relative to another is given by 100 e^/b in percent. Another
23
-------
form of this precision is given by the root mean square deviation concentra-
tion of the y-sampler from the regression fit y = a + bx. It is useful to
see this value for all samplers compared to the composite sampler. Finally,
it is useful to examine the extent to which the variables y and x are corre-
lated. This is revealed by the correlation coefficient. In this study all
samplers attempted to measure the same pollutant concentration, hence their
correlation with the composite mean will usually be near 1.0. However, when-
ever sampler malfunctions occurred and/or few results were reported, the re-
ported results will correlate poorly with the composite mean. The equations
used to calculate all of these quantities are given in Appendix B.
SAMPLER RESULT TABLES
The complete set of results from the study is given in Appendix C. Con-
centration values were reported for 13 different pollutant species in 3 size
fractions, the total, the small or fine, and the large or coarse particulate
fraction (39 Tables). The sulfur concentration results were multiplied by 3
and combined with the sulfate results to give 3 additional tables. Figure 6
illustrates a portion of one of the 42 tables (small sulfur).
The top most grouping lists the reported concentrations in appropriate
units as a function of period (1 through 16) of the study. The alphabetical
sampler identification letter is that of Table 4. A dash means no value re-
ported and a concentration value preceeded by a minus sign indicates an out-
lier value according to the Dixon Criterion (ot = 4%). The composite sampler
concentration or Z-mean is listed next along with its standard deviation. Any
sampler excluded is noted parenthetically, and as discussed above, all outlier
values are included in the Z-mean (except for sampler Q, small sulfur). The
X-mean and its standard deviation exclude selected duplicate samplers. Those
included in this calculation are shown in parentheses.
The next grouping lists all sampler concentrations normalized to the com-
posite Z-mean concentrations. Most of these entries generally lie near 1.0.
To the far right, the 16 period average ratio, its standard deviation, and the
coefficient of variation (percent error) are shown. An asterisk preceeding
any tabular values indicates that the participants reported error for that
period allows his value to overlap the Z-mean (1.0). If a participant's value
deviates far from 1.0 yet still has an asterisk, this indicates a rather large
error; while values close to 1.0 without asterisks may suggest a small error.
Below this grouping appears the mean ratios which by definition are always
1.0. However, its standard deviation reflects how tightly clustered all of
the reported results are as a function of period. Note that the 16-period
average standard deviation multiplied by the average pollutant concentration
yields the error given to the far right of the top most grouping.
The lower grouping tabulates results of the regression analysis of each
sampler vs the composite sampler (Z-mean). The quantities listed left to
right are the intercept and its error in either /^g or ng/m , the slope and
its error, the root mean square deviation in concentration units, and the
correlation coefficient.
Appendix C of this report contains only the regression analysis results
24
-------
RESULTS
INTERCOMPARISON STUDY C
PERIOD 1234
PERIODS 1-16, MAY 11-18 1977.
II 12 13 14 15
15
AVG SD
D EPA DZUBAY MDS-A
M EPA DZUBAY TF-M
N EPA DZUBAY 1F-P
R EPA DZUBAY MDS-S
C LBL LOO DICOT
L LBL LOO DICOT
S LBL LOO DICOT
G UCD CAM ILL SMDAY
I UCD CAHILL SFU
Q STL DEL/EO 2MASS
Z ME AN (WITHOUT Q)
STANDARD DEVIATION
X MEAN(D,M,L.G, I )
STANDARD DEVIATION
RATIO TO 2
0 EPA
M EPA
N EPA
R EPA
C LBL
L LBL
S LBL
G UCD
I UCD
Q STL
DZUBAY
DZUBAY
DZUBAY
DZUBAY
LOO
LOO
LOO
CAHILL
CAHILL
DEL /ED
ME AN I WITHOUT
PERIOD
MDS-A
TF-M
TF-P
MDS-S
DICOT
DICOT
DICOT
SMDAY
SFU
2MASS
0)
STANDARD DEVIATION
REGRESSION VS Z
D EPA
M EPA
N EPA
R EPA
C LBL
L LBL
S LBL
G UCD
I UCD
Q STI
DZUBAY
DZUBAY
DZUBAY
OZUBAY
LOO
LOO
LOO
CAHILL
CAHILL
DCL/ED
MDS-A
TF-M
TF-P
MOS-S
DICOT
DICOT
DICOT
SMDAY
SFU
2MASS
456? 2668 3737 2943 t
5009 3572 3894 4182 61
4951 3170 3443 3684 55c.
4297 3040 3707 3373 574.
5336 - - 4123 6572
5308 3518 4309 3945 6430
5244 3693 4274 - 6448
4310 2500 2940 3800 5570
3769 2693 3491 3481 7557
1900 1210 1270 -970 -1430
4754 3107 3724 3691 6157
548 458 451 416 670
4592 2990 3674 3670 6225
601 512 507 479 829
12345
•0.96 0.86 •! .00 0.00 0.90
•1 .05 1.15 »1 .05 1.13 "0.98
•1 .04 •! .02 0.92 • I .00 0.90
0.90 '0.98 •! .00 0.91 0.93
1.12 - - 1 . 12 M .07
1 . 12 1.13 1 . 16 M .07 M .04
•1.10 1 . 19 1.15 - M .05
•0.91 0.81 0.79 •! .03 »0.90
0.79 0.87 *0.94 '0.94 1 .23
0.40 0.39 0.34 0.26 0.23
1 .00 1 .00 1 .00 1 .00 1 .00
0. 12 0. 15 0. 12 0.11 0.11
INTERCEPT ERROR SLOPE
-219.166 396.721 0.963
387.016 195.029 0.944
62.634 207.518 0.934
-95.827 167.455 0.962
266.537 189.194 1.040
195.124 114.091 1.051
325.003 196.595 1.023
-1002.408 731.934 1.170
449.415 610.244 0.866
0. 0. 0.
•42 4336 7912 - 8951 - D
18 4695 7586 9298 6506 8687 M
4525 7764 - 8612 - N
4683 7362 - 9234 - R
5571 8475 9532 10125 9019 C
i468 8617 9851 9998 8969 L
i|4 - 9577 10146 9094 S
>. - - - - G
721 •? 6950 9314 7375 5833 1
-1151. -2090 -1390 -1920 3130 Q
7650 7809 9514 9131 8320 5869 626
603 "591 226 959 1399
7851 741 8 9488 8733 7830
423 45. 315 1081 1735
•1
0
0
•1
•1
•1
•1
•0
0
1
0
6
-
.03
.88
.91
.08
.08
.03
.04
.95
. 15
.00
.08
7
•0.97
•1 .01
•0.98
•0.96
•1 .08
•1 .08
•1 .07
•0.92
•0.95
0. 16
1 .00
0.06
14 15
0
- «0.98
18 0 . 94 • 1 ,
0.9
0.91
-
- •
- •
1 .31 •
•1.11 •
0. 14
1 .00
0.22
ERROR RMS.DEV.
D.
0,
0
0,
0.
0.
0.
0,
0.
0.
.070
.031
.036
.059
029
.018
,031
. 139
097
530.327
284 . 1 04
281.747
227.354
840.356
166.201
270.438
880 . 238
890 . 555
0.
- 0.94
•I .01
•1.11 *1 .
1.10 M .
1 . 11 •! .
0..
1.0 M 0.
0.23 0.
1 .00 l
0.08 0 i.
COR.COEF
0.97c
0.993
0.991
0.995
0.996
0.998
0.995
0.942
0.923
0.
1 6 AVG SD
-DO
, 04 M 1
-NO
- R 0
, 08 C 1
. 08 L 1
, 09 S 1
-GO
70 I 0
. 38 Q 0
00 1
17 0
.92 0.08
.02 0.06
.95 0.05
.94 0.03
.09 0.03
.09 0.03
.09 0.05
.94 0.15
.95 0.13
.29 0. 14
.00
. 1 1
CV
9
6
5
4
3
3
5
16
14
47
Figure 6. A portion of one of the tables listing results for the pollutant
small sulfur. Complete results for all pollutants can be found
in Appendix C. See text for an explanation of the entries.
25
-------
of each sampler with the composite sampler. Regression analysis results on
all pairs of samplers are available from the study director, the EPA project
officers, or any of the participants.
26
-------
SECTION 6
SELECTED RESULTS AND DISCUSSION
PRECAUTIONS IN INTERPRETING RESULTS
The objective of the study was to evaluate sampler performance by inter-
comparing the reported concentrations of selected pollutant species as mea-
sured by the samplers. The measured concentrations contain many factors which
influence their magnitude. Before presenting the results, it is worthwhile
examining some factors that can influence the concentration values obtained.
In Section 1, Table 4, the various sampler performance parameters were
presented. Among the more prominent differences were sampler flow rates, in-
let designs, particle size cut-points, and collection substrates. Those sam-
plers which had cylindrically symmetric inlet designs should be insensitive to
wind direction, but not necessarily wind speed. Samplers with such inlets
will accept particles up to some limiting aerodynamic size. Others which do
not have cylindrically symmetric inlets will be wind direction dependent. In
comparing total pollutant specie concentrations, the maximum size particles
accepted may influence the final concentrations obtained.
Some of the samplers brought to the study separated the incoming air par-
ticles into two or more size fractions. In comparing results from these sam-
plers, the particle size distribution will influence the results obtained. A
sampler having a fine particle cut-point of 2.5 micrometers cannot obtain the
same concentration as one having a 3.5 micrometer cut-point, all other factors
being equal. In comparing results for large particle sulfur, for example, a
sampler having a cut-point of 2.5 micrometers may record significantly more
large particle sulfur than one having a cutpoint of 4.3 micrometers particu-
larly if very little sulfur is associated with particles having diameters
greater than 3.0 micrometers. Also some filter substrate materials, particu-
larly fiberglass, will allow the formation of stable S04 ions from SO^ gas
which passes through them. This adds some mass and causes an artificial in-
crease in the SO^ concentration.
Another important factor is the analytical technique, its sensitivity,
accuracy, and precision. Supposedly all analytical techniques used are cali-
brated as carefully as possible. But different calibrations can result even
for identical instruments if two different "standards" are used. Thus, even
analytical technique calibrations can lead to differences as large as 5%, and
possibly 10% depending on the technique.
Clearly, sampler flow rates, cut-points, and analytical techniques can
lead to differences in reported concentrations. Other important parameters
27
-------
that may cause differences are the extent of particle bounce-off from sampler
walls, sticky vs non-sticky substrates, different methods employed in the ini-
tial calibration of a sampler flow rate and cut points, different methods of
monitoring sampler flow rates at the field study, screened or unscreened in-
lets, and possible inhomogeneity in the air actually sampled. These are just
a few of the sampler parameters which can lead to differences in the final re-
sults. Some of the differences may be explained and removed, other parameters
will be unknown and the differences they cause in the results not removable.
Therefore, the reader is urged to consider very carefully those factors which
can contribute to observed differences, before drawing conclusions about a
samplers relatively "good" or "bad" performance.
POLLUTANT CONCENTRATION VS PERIOD
Pollutant concentrations were reported for each 12 hour period injug/m
for all mass fractions and in ng/m for all other pollutants. Table 7 lists
the 16 period pollutant concentration averages and standard deviations for
all thirteen species measured. For the total fraction, maximum and minimum
values and their standard deviations observed during the 16 periods are also
given. For the small and large fractions, only the 16-period averages and
standard deviations are tabulated. The sum of the latter two averages is not
always equivalent to the total average because some sampler types measured
the total fraction, but not either sub-fraction. An inspection of the results
shows that large particulate crustal elements have the largest standard devia-
tions. On the other hand, those elements typically found in the small parti-
culate fraction are somewhat better measured even though their absolute
concentration is less.
Not all of the various pollutant species measured behaved in the same
manner throughout the study. Figure 7 shows several selected pollutant specie
composite sampler concentrations (the Z-means) as a function of half day peri-
ods for the total and small fractions. The error bars shown represent stand-
ard deviations. The slightly displaced squares indicate the small or fine
fraction (< 3.0 + 0.5jim) Z-mean concentrations. From Table 4 it was clear
that there was a wide range in particle sizes selected by the sampler inlets
(as well as the fine particle cut-points), hence part of the spread in the
standard deviations is due to variable size particles accepted by the differ-
ent sampler inlet designs and cutpoints. The standard deviations in the
crustal elements, silicon, calcium, titanium, and iron, show the greatest
variation; mass shows somewhat less variation. Pollutants typically concen-
trated in the fine fraction show the smallest standard deviations.
Throughout Figure 7 the odd periods represent daytime collections (0800 -
2000 hrs) while the even periods represent nighttime collections (2000 - 0800
hrs). Except for sulfur the lowest pollutant concentrations were recorded
during period 9. Note that the silicon and iron concentrations track each
other closely, and mass behaves similarly to the crustals, but is not identi-
cal. The zinc concentrations were approximately constant except for two large
excursions during periods 2 and 4. Lead concentrations peaked Friday and
Saturday nights, periods 6 and 8. Sulfur slowly increases as the study pro-
gressed except for the 24 hour dip beginning Sunday at 0800. Visibility was
good during the beginning of the study, but deteriorated steadily ending in
28
-------
TABLE 7. 16-PERIOD POLLUTANT CONCENTRATION AVERAGES FROM ALL
SAMPLERS REPORTING RESULTS FOR MAY 11-19, 1977
Pollutant
Mass3
Ni trate
ng/m
Sulfate
ng/m
Sulfur
ng/m
Silicon
ng/m
Calcium
ng/m
Titanium
ng/m
Iron,
ng/mj
Copper
ng/m
Zinc,
ng/m
Selenium
ng/m
Bromine
ng/m
Lead,
ng/m
Total Fraction Small Fraction
Max & Min
Avg Values Recorded Avg
146 + 23
104 + 18* 55 + 13
47 + 6
2860 + 2400
1880 + 1000 " 1020 + 760
860 + 450
30600 + 3800
19350 + 3000 18200 + 3000
6420 + 2360
10800 + 970
6640 + 1000 5870 + 630
1935 + 400
17600 + 7200
9950 + 3650 1140 + 520
4000 + 2100
4200 + 1500
2500 + 900 300 + 160
730 + 350
400 + 100
250 +90 " 44+29
88 + 46
2400 + 670
1750 + 500 " 350 + 150
730 + 240
400 + 70
88 +40 80 + 30
24 + 12
283 + 41
116 +22 " 76+13
62 + 14
27+4
10+3 " 9+3
3 + 1
489 + 126
236 + 59 190 + 48
65 + 19
1660 + 260
1070 + 180 " 820 + 90
326 + 93
Large Fraction
Avg
43 + 15
1090 + 500
930 + 400
574 + 240
7680 + 3300
2230 + 1000
210 + 100
1400 + 600
17+9
40 + 13
1 + 1
64 + 23
220 + 90
* Standard deviations
29
-------
an air stagnation alert (throughout the southeastern mountain states) that be-
gan 36 hours before the end of the study.
Typically, the small mass fraction constituted about half of the total
mass collected; while small silicon and iron constitute only 11 and 20% of
their totals, respectively. About 80% of the sulfur and lead are found in the
small fraction, while zinc is almost equally distributed 60/40% in the small/
large fraction. All of the results from which these plots were made can be
found in the complete tabulation of the total, small and large fraction re-
sults in Appendix C.
AIR HOMOGENEITY
As mentioned in Section 1 all samplers were arranged in a linear array
1.5 meters from the southeast face of the building. Sampler A was located
about 9 meters southwest of the eastern corner of the roof, and sampler U a-
bout 9 meters from the southern corner of the building. Two CHAMP samplers,
B and U, were duplicates and collected both large and small fractions; sam-
plers C, L and S were the automated dichotomous samplers (ADS); and samplers
D and R were the manual dichotomous samplers (MDS). Note that samplers B, C,
and D were adjacent while R, S, and U were nearly adjacent. The former group
was toward the eastern end; the latter group toward the southern end. Sampler
L was located nearly centrally. The three automated dichotomous samplers C,
L, and S were brought to the study and used to evaluate the homogeneity of the
air sampled on the roof (see Section 4, page 13).
One method of evaluating any possible inhomogeneity in the air sampled is
to ratio concentration results from pairs of duplicate samplers. This can be
done for selected pollutant species, say mass and the crustal elements in one
group, and man-made elements in a second group. Table 8 lists the ratio of
pollutant concentrations as obtained from 4 pairs of samplers for three size
fractions and two pollutant groups. Ratios obtained from the two ADS pairs
for the crustals pollutants in the total and large fraction suggest that about
20% more crustals are collected at the south end relative to the east end.
However, the MDS and CHAMP sampler pairs do not observe this dispersion. For
the small particulate fraction in either pollutant group there is no disper-
sion. The MDS samplers suggest a reverse dispersion for the fine particle
crustal elements at the east end. For pollutant group B (anthropogenic trace
elements) there really is no strong evidence for any south to east dispersion.
In retrospect it would have been useful to have operated the ADS samplers side
by side before or after the study to verify that they would yield identical
results. However, such tests were not carried out.
Results obtained for wind velocity and direction, which were monitored
atop the Federal Building, indicated a strong diurnal wind pattern. Winds
were predominantly from 300° + 45° during the day, while at night they were
from 120° + 15°. If the south to east dispersion were caused by the roof
penthouses one might expect it to disappear or reverse as the winds changed
directions from night to day. Ratios similar to those in Table 8 were ob-
tained for the same pairs of samplers, only grouped according to odd and even
periods, and mean values obtained. There was no significant statistical evi-
dence for any diurnal variation of the dispersion. That is, the ADS sampler
30
-------
00 CO
CU
i— -0
( 1 ,-
1 1 1
— 1 O 1 -n-
_ i — o i -Q-
1 ft I n
- i— o— i -n-
1 O I n
0 i—ft—ld
U 1 O 1 "•
~c75 ' ° '
— i Q | n
i ft I '~n~
1 Q i n
n
— | o 1 -a-
1 0 1 n
1 O 1 "^
1 1 1
1 1 1
_o
rtj
Q_
1 ' ' "KH
0 1 HDH
- !— 0 — 1 *^-
- I o 1 hDH»
_ |_o — [ ^^y.
1 rt 1 rO—t _,
r ^^ i
|O|
0 M>ln
V— i Q_ i H-n
0 1 I~CH
( \J 1
— 1 O 1 H3H_
i o 1 ha~l
1 w 1
, 0 | KH
1 O 1
— 1 O 1 ^~
, o | hCH
I 1 1 1
111!
I
"O
_o
'i—
a.
11^ 1 1
— otSr1
— Of 1
t-o-P —
. H-oifi,
_ )_$$
| ^Ofl
-£ «fl-
— a
' ^ ,_g |
1 — O — "
- (-0^
— (-O^i1 —
LHJ-J
- KJ* -
KS1
— hO^1 —
._.,
II II
II II
S— (T3
2 'o "
JO
a.
^" re "
^ +-> a>
o u_
CM ^
"~ S- C
O hvl
o ^
"D 4J CO
O r—
00 QJ to to
Q- 0) to
s- re
to o, E
c to
-P C
•^ s- re
0 4->
0. 3
a> i —
(M i- r-
O
CO f^
s-
aj fli
3 o ooooooooo r^-£
r> o ooo ooo g--^
3 O OOO OOO =.
D 0 <*> ™ ^ 00 •* « ^
NT— r— 4_
J918LU Ojqno/siUBjBouBN J8J91U o!qno/soiBj6ouB|\| J818W ojq no/soi BjBojoiiAi ^ £
o
0 -LJ
I 1
' 1 ft 1 '
u_ft_i r»-*i
^^ r^w^^ —
- 1 — 0 1 1— O— 1
l_f-L_l
— I-OH h"O~l —
1 •-« 1
_ i -^V..-.. J r \J" — " _
— HO— 1 h~DH —
i j* i |— -D — I _
I \^ ,
-s lot01 -
CO
— S hOH WH —
H-OH KH
._, (CH
r—^—f
! o , 1— 0-H _
HCM »« -
1—0— ( HS-1 -
_ , o | K3H
h-0— t H3H
1 1 1
T3
g
O)
Q.
1 —
1 1 1 1
i— KX D —
u HOH 1QI -
LnJ
— KH "^ —
i- t-ft-i IDI —
^ n\^n
— KH D —
— (01 a—
"~ ^ hOH
— N i — tio^s^r1 —
MH D -
— hOH D
I-CH tH
.. i o- ^r0"1 —
h-0— tp^
o i — p— i —
|_0_|H3l -
1 1 1 1
T3
o
aj
Q.
| |
1-0— ln
— 1— O-'-P"^ —
,__0_hpH
_ l-OtfH —
- hO^ _
TO
__1 n**y^
i_ ,Q_ _ fOt _
-t— o — h0-1 —
I-OHK"
_ H-o-irCH -
(-O-lpt
_ i-o-tp4 -
l-O-ifH
1 1
O
CD to re
*~ 04-
•r*
^ 4-> CU
T— re N
i- -I-
-I-) tO
CM E
r— d) CU
U i—
C O
^ "§ «= re
.2 re a.
S- CD
10 0 S
C" O
CNJ
c re
0 E
CN CU to
X T3
•r- C
LO (O
D O OOOOOO O O
D O O O O o O
NT— ro CN <— o o
CN T- rx
en
31
-------
showed the south to east dispersion consistently for pollutant group A, total
and large fractions, night or day; while the MDS did not. The observed dis-
persion, then, appears to be independent of any micrometeorological wind
streams associated with the penthouses or roof.
TABLE 8. POSSIBLE SOUTH TO EAST DISPERSION IN THE AIR SAMPLED
Pollutant
Group
a)
Total
Fraction
A
B
S/C
ADS
1.19 +
1.06 +
Sampler Identification and Ratios
S~7TR7D D7B~
.05
.05
b)
ADS
1.08 +
1.04 +
.02
.05
MDS
0.99 + .05
0.97 + .12
CHAMP
0.98
c)
Large
Fraction
A
B
1.22 + .01
1.10 + .08
1.08 + .02
1.02 + .08
1.04 + .05 . 1.05
0.97 + .09 d'
c)
Small
Fraction
A
B
1.00 +
1.03 +
.05
.05
0.98 + .04
1.04 + .06
0.74 +
0.98 +
.06 e) 0.96 c)
.08
a) Pollutant Group A contains mass, silicon, calcium, titanium, iron,
i.e., the crustal elements. Pollutant Group B contains the anthro-
pogenic elements sulfur, zinc, bromine, and lead.
b) Standard deviation.
c) Only mass was determined by the CHAMP samplers.
d) Excludes the large zinc fraction ratio (0.73).
e) Excludes the small mass ratio (1.04).
ANALYTICAL TECHNIQUE QUALITY CONTROL SAMPLES
From one of the automated dichotomous samplers operated by LBL, a set of
30 fine fraction samples was selected. These samples were analyzed by two
referee analysts (Robert Giauque of LBL and Norman Bonner of LLL) for Ca, Ti,
Fe, Cu, Zn, Se, Br, and Pb. Silicon was not determined because of significant
particle size corrections that must be applied for the soft silicon x-ray.
Agreement between the results from the two referee analysts was good. Regres-
sion analyses were peformed on the two sets of data and the results are tabu-
lated in Table 9. In two past intercomparison studies (6,7) in which these
two referee analysts have participated the LLL values were also observed to be
5X or so lower than the LBL values. Thus, for this study adopted concentra-
tion results were used which were essentially mean values of the reported
referee results.
32
-------
TABLE 9. REGRESSION ANALYSIS FITS OF REFEREE ANALYSTS RESULTS
Element
Ca
Ti
Fe
Cu
Zn
Se
Br
Pb
LLL = (B + eb)LBL
0.915 + 0.071
0.933 + 0.114
1.003 + 0.026
0.986 + 0.012
0.937 ± 0.007
0.887 + 0.018
1.063 + 0.020
0.935 + 0.014
+ (A + 6A)
66.2 + 25.9
8.2 + 7.7
-12.0 + 15.0
-4.6 + 4.9
9.2 + 2.1
1.2 + 0.7
-6.3 + 9.2
8.9 + 36.2
No. Values
LLL/LBL
30/30
29/23
30/30
29/30
30/30
28/28
30/30
30/30
Mean Slope 0.957 + 0.056
Sulfur was analyzed on all 30 samples by one of the participants (Loo)
and on 6 samples by one of the referee analysts (Giauque). Before the sulfur
measurements could be performed by the referee the six filter samples had to
be cut down to 2.54 cm diameter circles. Since the sulfur analyses were per-
formed after the samples were cut, there was a possibility that material may
have been lost as a result of the cutting process in which case the sulfur
data would be biased. Five of the six filter samples were analyzed for sul-
fur, and then reanalyzed at LLL for the 8 trace elements previously measured.
The sixth sample was rendered unmeasurable in trying to cut it. Table 10
shows the results for the LLL reanalysis. Sample 80721 may have been conta-
minated somehow by a copper-zinc source because the values were unquestion-
ably different afterwards. The selenium concentration on sample 80730 was
very low (large error). Otherwise there is excellent agreement before and
after cutting. There is evidence that the bromine concentration changed and
additional evidence for its loss with time is discussed below. There was ex-
cellent agreement between Loo's 5-sulfur results and the LBL referee values,
hence Loo's sulfur values were used as reference values for the remaining 24
samples.
33
-------
TABLE 10. RATIO OF XRF RESULTS FOR CUT TO UNCUT FILTER SAMPLES
Cut Sample Concentrations (ng/cm_) by XftFA
Uncut Sample Mean Concentrations (ng/cm) by XRFA
Sampl e
Number
80721
80724
80725
80730
80734
MEAN
Ca
1.053
+.167a)
0.966
+.126
1.062
+.131
1.000
+.168
0.993
+ .157
1.015
+ .041
Ti
N.D.
0.903
+.169
1.288
+ .264
0.630
+.226
1.188
+.223
1.002
+ .297
Fe
0.959
+.163
0.897
+.125
1.125
+.138
1.126
+.157
0.943
+.151
1.010
+.108
Cu
1.809
+.177
1.092
+.201
1.081
+.159
1.051
+.150
— — _
1.223
+.394
Zn
1.368
+.191
0.879
+ .201
1.023
+.191
1.026
+.145
— — —
1.074
+.208
Se
N.D.
0.861
+ .150
1.000
+.264
0.571
+.566
N.D.
0.811
+.219
Br
0.814
+.203
0.819
+.120
0.912
+.159
0.945
+ .157
— _ —
0.873
+.066
Pb
0.935
+.142
0.933
+.142
1.052
+.141
1.036
+.141
0.896
+ .141
0.970
+.069
GRAND MEAN
+error/#
samples
1.003
+.209/34
WITHOUT [ ] Entries
1.028
+.068
0.976
+ .084
0.931
+ .098
0.979
+.123/31
a) Standard deviation
Eight sets of three samples each were mailed to the remaining partici-
pants. Some of the participants did not make trace element concentration
measurements on their own samples, so were not required to measure them on the
reference samples. The results for all elements except silicon are shown in
Table 11. Except for the Washington University group these results eliminate
the analytical technique as a source for any large (> 10%) differences in re-
ported pollutant concentrations. The bromine values appear to be unstable
with time and will be discussed more completely below.
34
-------
TABLE 11. PARTICIPANT RESULTS FROM ANALYTICAL
TECHNIQUE QUALITY CONTROL SAMPLES
Ca Ti Fe Cu Zn Se Br Pb Mean
Cahill a)
UCD (3)
Dzubay
EPA (3)
Loo (30)d
LBL
Hudson
FSU (3)
Rodes
EPA (2)
Tanner
BNL (3)
Del umyea
WU (3)
1.01
+.07
0.93
+ .04
0.98d)
+.04
0.89
+ .05
1.06
+.16
0.84
+.12
0.76
+.15
1.
+.
0.
+.
1.
+.
1.
+.
—
41 b>
26
89
06
27
13
06
16
1.04
+ .43
1.14
+ .13
1.33
+.19
1.01
+ .05
,^
1.01
+ .06
0.96
+.05
0.98
+.05
0.94
+.04
_
0.90
+.01
1.04
+.13
1.02
+.07
1.33
+.44
._
1.14
+ .05
0.95
+.06
1.10
+.06
1.12
+.20
_
0.93
+ .06
0.96
+.14
0.87
+.14
(f)
_
0.91
+.05
0.92
+.05
1.16e)
+.06
0.92
+.07
_.
0.92
+.03
0.93
+.02
1.03
+.02
0.84
+.05
1.03
+.02
0.98 c)
+.08
0.97
+.08
1.08
+.15
1.01
+.16
_
Burton (EPA) and Mueller (ERT) did not report trace element concentrations.
a) The number of reference samples measured is given in parenthesis.
b) The filters were temporarily exposed to a known source of calcium contami
nation (cement dust).
c) Mean value calculated without including calcium.
d) All 30 reference samples were originally from one of the automated dicho-
tomous samplers. Only five could be compared with the referee results.
e) These bromine values were obtained early, then divided by the referees'
results which were obtained later after an apparent bromine loss. See
the Bromine Loss discussion below.
f) Selenium concentrations were to small to be accurately measured by PIXE.
AVERAGE SAMPLER RATIOS VS SAMPLER TYPE
Each sampler's 12-hour pollutant concentration was normalized to the Z-
means to yield ratios near 1.0. These ratios can be averaged for all the
periods to obtain an average sampler ratio for the duration of the study.
For example, if the high volume sampler consistently collects more total mass
than the composite sampler Z-mean, its mean ratio for the entire 16 periods
would be greater than 1.0. Another sampler with a more restrictive inlet
would have a ratio less than 1.0. An examination of these mean ratios can
show systematic trends, illustrate consistent differences between sampler
35
-------
types, and even give an indication of the accuracies and precisions for groups
of samplers brought to the study by an individual participant.
Figure 8 shows results for a composite of four selected pollutant
species. The 16-period average ratios are plotted as a function of sampler
type. The abscissa lists the sampler identification number (see Table 4).
The samplers are grouped by participants and arranged within each group in an
east to south direction. The top two plots display those samplers reporting
total and small mass. Some participants reported both. The two ordinary Hi-
Vol samplers P and T report the highest relative mass, undoubtedly because of
their ability to collect larger sized particles (unrestricted inlet size).
These two samplers report 50% more total mass than the two manual dichotomous
samplers, D and R. The CHAMP samplers, B and U, with inlet restrictions
report 15% less mass than P and T. Sampler E, the cyclone unit, also reports
relatively lower total mass. For small mass, all of the samplers excluding B
and U report an average mean sampler ratio of 0.89 + 0.03 vs 1.36 + 0.03 for
B and U, a factor of 1.56 times as much mass. This is a large difference and
suggests that more than a single sampler performance parameter is responsi-
ble. Sampler 0 (Rodes manual dichotomous reports more variability than do
samplers D or R (Dzubay manual dichotomous). Sampler 0 is an older model.
The middle set of plots display all of the sampler results for total sil-
icon and iron. Note the much larger scatter in these two elements compared to
total mass, the larger standard deviations, and the same patterns for groups
of samplers for the two elements. Sampler H is anomalously high only for sil-
icon, not for iron (or calcium or titanium). The systematic increase shown by
samplers C, L and S - east to south is the same 20% air inhomogeneity dis-
cussed above, but it is not evidenced by the ratio of samplers R to D, or U to
B (total and small mass). The two tandem filter samplers M and N also record
more iron (and titanium and calcium) than silicon, relative to D and R. These
two tandem filter samplers operate in a similar manner to the UCD stack filter
sampler, I, but record relatively more iron than silicon.
The two lower plots display the behavior of the mean sampler ratios for
total and small lead. The tight clustering of the small lead results and the
generally smaller standard deviations for both is immediately apparent. Note
the lack of statistically significant evidence in samplers C, L and S for the
south to east air inhomogeneity. As discussed above the crustal elements ex-
hibited such an inhomogeneity in these samplers, but the fine particles, e.g.
sulfur, lead, and zinc, did not. Once again, the pattern of certain groups of
samplers high, others low, is repeated. Thus, a given participant has repro-
ducible results regardless of which kind or type of sampler he operates, but
apparently there are real group to group calibration differences. For example,
LBL (samplers C, L, and S) reported 10% more lead in the analytical technique
control samplers (Table 11) than did EPA (Dzubay-samplers D, M, N, and R) or
UCD (samplers G, H and I). If an appropriate correction is made to their
small lead average ratios averages there is even better agreement in the over-
all results. However, how does one correct for the fact that the fine parti-
cle cut-points of samplers C, L, S, and I are about 2.5 micrometers, versus
3.5 micrometers or greater for samplers D, M, N, R and G? The former should
record less small lead, not the same, unless there are yet other compensating
36
-------
O I
1 i i i
— i-
— ^-O~^
— H-O— i
—
cn
cn
03
- E
— "CD
E
— CO
1 i i i
i i i l
O, i
O— i —
01
K5-I —
KH —
(OH —
HOH —
K-O-I —
1 1 1 1
a
U-
D
03
O
CO
_J
o
DC
Q
N <- in r
o
i r i i
— t-O
— K3H
— I-O-H
— K>
— cn tO
CO
cn
— c
— ~ca
JT->
O
1 —
', 1 1 1
1 1 1 1
t-OH —
1 —
—
—
-
cx —
I-O-H —
KH —
1 l l i
LU
D
1-
QQ
CL
CO
_1
O
cc
Q
1 I 1 1
— 1— O — 1
— 1— O— 1
CU
— LL 1 O-
— 2 )-Oi
o
- 1-
i 1 1 1
i i i i
— i —
1— 0— l —
1 0 H-
I-OH —
— 1 —
— 1 —
HO-I -
1 I 1 1
^
—)
1
CD
CO
_l
o
DC
-z.
^
Q
I'll
— KW
_ lOt
-£ ^
— "CD >-O
— CO >
1 I i i
1 1 1 T
-OH _
-OH —
KM —
H —
H —
X —
1 1 1 l
vl <— LO CN «- u
0 C
(TIT
— 1_ /v- i
- ^ i
- I-O-H
- ,-0-
- w L
n:
-M
o
i i i i
ill i
—
—
I-O-H —
K>H —
1 1 1 1
^
—j
X
CD
CO
_J
0
I
— KH
JD
- S H
O
1 i 1 1
i i l i
K>H —
l-OH
—
—
KX —
v_j
KH ~
i i i i
O
o
CO
_i
O
DC
2
^
Q
cn
L-
cu
T Q.
3 £
03
CO
^
—5
a_
X
a
CO
_l
o
DC
Z
2
Q
^ «— uncN <— IDCM <— in
odd
CL E
13
OJ
S- -E
o 4->
i+-
s_
to (D •
CD
S-
CD
to "O .E
CD CD
i- cn E
I
i— 0
to
SOI;BJ
03
E S-
O OJ
•r- >
-t-> 03
O3 to
S- -O E
-t-> E O
E 03 •!-
O) -l->
O E 03
E O3 T-
O O) >
O E OJ
i ~a
•O l^vl
OJ TD
4-> CD S-
S E O3
O 4-> -O
CL E
Ol O 03
S- 4-> 4->
to
s- -o
CD
S -E
4->
CD 4->
CLr— -O
>> =5r-
4-> tO CD
O) T-
s- s- >,
CD
•— T3 O
Q- O) 4J
E 4J
03 S- -O
to O CD
CL4->
to CD S-
> S- O
CL
O JC CD
•r- O S-
4-> 03
03 LU to
CD • rs
CD tO to
03 -O O)
S- O S-
ai -i-
> s- <+-
O3 <1J O
Q.
a; s-
-E 1£> 0)
l— -—I .a
CO
ID
s-
CT)
37
-------
factors. For total lead concentrations inlet particle size restriction dif-
ferences may be invoked to explain any remaining differences. Sampler 0,
which operates on the same principle and has the same flow rate as D and R,
is less precise. It also shows a larger variation for small mass than did
D and R.
Figure 8 is an example of the kind of comparison that can be drawn for
the pollutant species measured at the intercomparison study. The foregoing
discussion points out some of the striking differences, and in several cases,
advances possible explanations. Other comparisons of sampler average ratios
will be discussed below.
PAIRWISE PRECISION OF SAMPLERS
There were 6 participants who brought two or more samplers of one type
to the field study. Of these 4 reported results that could be intercompared.
Since one participant brought triplicates and another, two different dupli-
cates, 7 sets of samplers can be pairwise compared. One measure of the pre-
cision of one sampler -ersus another is contained in the error on the slope
of the linear regression fit to the results from both samplers. Since a re-
gression analysis was performed on all pairs of samplers, the data are avail-
able. In the expressions
y = a + bx or x = c + dy (4)
the percent error or precision, P, is given by
P = 100 eb/b = 100 ed/d (5)
where e is the error in the slope b or d. Note that, in general, b j- d, nor
does eb = ed, unless the correlation coefficient and both slopes are nearly
exactly equal to 1.0.
Table 12 lists the precisions of the 7 pairs of samplers for selected
pollutant species in the three size fractions. Of course, not all of the sam-
plers were able to measure each of the size fractions hence the blank entries.
For example, samplers J and K measured only the total fractions. The best
precisions were obtained on fine sulfur and lead as was suggested by the
standard deviations shown in Figs. 8 and 9. For these two pollutants, the
automated dichotomous samplers obtained excellent precision or reproducability
(1.4 to 2.3%). When the good to excellent precision obtained by the manual
and automated dichotomous samplers for the fine particulate fraction is com-
pared to results for the large fraction, there is a suggestion that large
particulates are not as easy or reproducible to collect. In addition, it
appears that the automated dichotomous sampler, S, did not behave as uniformly
as did C or L. Perhaps there are variable amounts of particle bounce off or
other loss mechanisms associated with large particles in dichotomous samplers;
or differences between sampler inlet designs.
Appendix C contains all of the reported results for all pollutants, par-
ticipants, and periods. However, the only linear regression analysis results
included are those for each sampler versus the composite sampler (Z-mean).
38
-------
Detailed regression analysis results for all pairs of samplers can be obtained
from the study director, one of the EPA project officers, on any of the parti-
cipants.
TABLE 12. PAIRWISE INTERCOMPARISON OF SAMPLER PRECISIONS (%)
Frac-
tion
T
0
T
A
L
S
M
A
L
L
L
A
R
G
E
a) The
b) The
c) The
d) The
e) The
0.2
Pollu- 3urtona) Dzubayb) Dzubayc) Lood) Lood)
tant B vs U M vs N D vs
MASS 6.2 10.6
SULFUR 7.3 5.9
IRON 14.3 8.2
LEAD 5.7 5.1
MASS 4.8 7.0
SULFUR 5.0 6.1
LEAD 5.3 3.3
MASS 15.2 18.5
IRON 13.1 9.3
ZINC 14.5
R C vs L L vs
4.6 8.9
1.5 1.8
9.0 15.6
1.7 2.7
2.5 3.4
1.5 1.9
1.7 1.4
8.0 13.1
9.2 17.8
8.0 30.0
CHAMP sampler with inlet size restrictions and one
tandem filter units; however, di
manual dichotomous samplers.
automated dichotomous samplers.
linear streaker samplers, one wi
H nucl eopore substrate filters.
fferent substrates
th 0.4 n nucl eopore
Lood) Hudsone)
S C vs S J vs K
8.3
2.3 5.7
11.8 12.6
2.8 11.2
4.1
1.9
2.3
11.0
11.8
14.2
size cut.
were used.
, the other with
SULFUR-SULFATE COMPARISONS
Recently it has been shown (8) that sulfate concentrations can be ob-
tained by determining the elemental sulfur concentration and multiplying by
three. Figure 9 shows an intercomparison of the sulfur and sulfate concentra-
tions as reported by the various participants. The sulfur concentrations were
multiplied by three, combined with the sulfate results, and a combined Z-mean
concentration determined. Then, the normalized ratios were calculated for
each period, and regression analyses were performed on the combined result
set. These results are included in Appendix C. The 16-period average sample
39
-------
2.0
1.5
1.0
0.5
I I I I
1.5
1.0
0.6
Total
-SO,
J I
I I
I I I
J I I
R PTAE
DMNRC LSGH
Sampler identification
i i r
so,
i r i
Small
I I T
—s
o 0
K
i i i i i i i
I
I I I I
I
0.38
i ±0.36
I 1 I I
ROBUFDMNRC LSG I (Q)
Sampler identification
Figure 9. 16-Period average sampler ratios for sulfur and sulfate combined.
It was assumed that the determined elemental sulfur concentration
times 3 is equivalent to determined sulfate concentration. Only
one sampler, R, reported both. Participant using sampler Q
reported results which were low.
40
-------
ratios are plotted vs sampler type for total and small sulfate-sulfur. Filled
symbols represent sulfate, open symbols sulfur. Only one participant (Dzubay-
EPA) measured both for the same sampler (R). It is assigned a square symbol
versus circles for all other samplers.
Sulfur occurs predominantly in the small fraction; and comparison with
the small lead results shown in Figure 8 reveals the same samplers are rela-
tively high and low. Samplers C, L, and S suggest no south to east inhomoge-
neity for small sulfur. Dzubay's group (D, M, N, and R) and the UCD group
(G, H, and I) report on an average 10% lower concentrations than the LBL
group. However for sulfur, Table 11 indicates that there is not a 10% dif-
ference in the analytical technique quality control results for these three
groups as there was for lead. Sampler G (the Sierra Multiday - UCD, fine cut
point < 4,3 jjm) reports only 5% as much sulfur as does sampler I (stacked
filter unit, UCD, fine cut point < 2.6 pm) in the large (coarse) particles.
Yet for total sulfur the former two groups are lower than the LBL values
by about 18%. Therefore, from the total sulfur and lead results there appears
to be at least a 10% difference between what samplers D, M, N, R, G, H, and
I collect and what C, L, and S collect. Samplers J and K, the linear
streakers, also report relatively more sulfur.
For the total sulfate results, the two EPA Hi-Volume samplers, P and T,
are 28% higher than the average of the three remaining samplers R, A, and E.
(Note that sampler A is a Hi-Volume sampler, which used phosphoric acid treat-
ed filters.) Similarly for small sulfate, the two samplers, B and U are 28%
higher than the average of the three remaining samplers R, 0, and F. This 23%
increase in both the total and fine fraction results may be due to artifact
formation (9) of SO* from S02 by the fiberglass filters used in samplers P, T,
B, and U. The standard deviations in the sulfate determinations do not appear
to be as good as those for total and small sulfur (+ 12% for SO^ vs + 8% for
S). If the sulfate to sulfur ratio is calculated from the total and small
concentration results instead of assuming it to be 3.0, the 16-period means
are 3.03 + 0.34 for total sulfate to sulfur, and 3.11 + 0.23 for the small
fraction. Thus the assumption of obtaining sulfate concentrations from ele-
mental sulfur results multiplied by three, is supported by results obtained
in this intercomparison study.
NITRATE RESULTS
An inspection of the reported results in Appendix C for total or small
nitrate concentrations reveals a number of strange variations in reported con-
centrations. For total nitrate the sampler pairs R and T showed higher ni-
trate concentrations relative to themselves during the day (odd periods) than
at night (even periods). The samplers A and E relative to themselves showed
the opposite pattern, that is, higher nitrate concentrations at night.
Nitrate concentrations in the small fraction were measured by samplers
R, B, U, and F; coarse nitrate was measured by only sampler R. Samplers E and
F operate similarly except for fine particle cutpoint. No diurnal pattern
similar to that recorded for the total fraction was evident for those samplers
reporting small nitrate. Except for periods 1, 14, and 16, sampler R reported
nitrate concentrations below their detectability limit of 50 ng/nr, compared
41
-------
to concentrations of nearly 4000 ng/m3 reported by sampler F for periods 3 and
8.
There are two hypotheses which may explain the wide disparity of nitrate
results. The results varied by a factor of 4 to 9 for total nitrate concen-
trations and up to two orders of magnitude for the fine fraction nitrate
concentrations. Neither hypothesis invokes problems with nitrate determina-
tion after collection and extraction since all investigators determined ni-
trate by either a reduction-colorimetric technique or by ion chromatography.
Cross-checks by both methods of the same extract from samples derived from
sampler A yielded roughly equivalent results: mean (1C [N0^]/colorimetric
) = 1.03 + 0.57. J
One hypothesis is that the differences in nitrate levels reported are due
to the presence of nitric acid vapor and/or N0£ in the sampled air, which was
collected with varying efficiency depending on the filter material used, and
analyzed as nitrate. Based on the work of Spicer (10) on nitric acid collec-
tion efficiency one would predict total nitrate values for the reporting
samplers (R, T, A, and E) which increased in the order R (Teflon) < A (acid-
treated quartz) < E (Teflon-coated glass) < T (glass fiber). The actual order
was R < T < A < E and it is necessary to introduce other explanations for the
observed nitrate concentration order.
A second explanation is that already-collected nitrate-containing parti-
cles may be impacted by acidic sulfate particles resulting in the topochemical
reaction
* + SO?" (4)
releasing nitrate from those filter materials which, by reason of surface neu-
trality do not adsorb nitric acid vapor. This might explain why, for example,
the total nitrate values of sampler A sometimes are close to those of sampler
R (periods 7, 13; acidic sulfate present) while at other times sampler A val-
ues approximate those of sampler E (periods 4 and 6). But this hypothesis
does not explain the relatively low results of sampler T. Thus, results from
this study would suggest that further extensive laboratory and field studies
are required before credible nitrate concentrations can be quoted for ambient
aerosols.
PARTICLE BOUNCE
Recently, John et al (11) described tests on the filtration of solid par-
ticles by Nuclepore filters in which evidence for the phenomenon of particle
bounce was presented for tandem or stacked filter units. The mechanism for
this phenomenon is pictured as solid particles nominally in the coarse frac-
tion rebounding from the filter surface, being re-entrained in the flow, pene-
trating the filter, and being collected as part of the fine fraction. By
contrast liquid particles stick to the filter surface inihi biting the bounce
phenomenon. Thus, the collection efficiency assumed for the coarse or fine
fraction may become dependent on the physical state of the aerosol particles
collected, and possibly on the substrate used.
42
-------
Results were reported by two groups that used either Tandem Filters (TF)
or Stacked Filter Units (SFU). These two samplers operate similarly; however,
their fine cut points were different, 3.5,1.1m for the TF and 2.6 ^m for the
SFU. Nevertheless, some evidence for the bounce of coarse particles into the
fine fraction can be found for these samplers if a comparison is drawn with
other samplers having similar fine cut points. The manual dichotomous sam-
plers had a 3.5 ^im fine cut point, while the automated dichotomous samplers
had a 2.4^m fine cut point. By ratioing the results reported for one sampler
to another for pollutants found predominently in the fine fraction, e.g., Zn,
Pb, and S; and then for those found primarily in the coarse fraction, e.g.,
Si, Ca, and Fe, a relative measure of bounce into the fine fraction can be
obtained. Table 13 lists the ratios of 16-period averages from just the
small fraction results for selected samplers. For example, samplers M and N
obtained 16-period average ratios of 1.13 and 1.03 respectively for Zn (see
results, page 132, Appendix C), while D and R obtained 0.99 and 0.88, respec-
tively. Ratioing the sum of (M + N) to (D + R) yields the ratio of 1.16 shown
for Zn.
TABLE 13. COARSE PARTICLE BOUNCE INTO THE FINE FRACTION
Selected Sampler Ratios of 16-Period Average Small Fraction Results
Element Element
% in Small
Fraction31
M+N
D+R
b)
c)
Zn
57
1.16
0.88
Pb
77
1.11
0.86
S
90
1.06
0.87
Mean
Fine
1.11
+.05
0.87
+.01
Si
11
1.45
2.54
Ca
12
2.53
1.70
Fe
20
2.10
1.44
Mean
Coarse
2.03
+.54
1.89
+.57
Mean C
Mean F
1.83
2.17
3(1)
C+L+S
The coarse particle content in the fine fraction for the
tandem or stacked filter samplers relative to the manual
or automated dichotomous samplers appears to be about 2.
2.00
+.17
a) From Table 14 (page 48).
b) M and N are the Tandem Filter Samplers (Dzubay) and D and R are the
Manual Dichotomous Samplers (Dzubay). The 16-period average ratios for
individual samplers and particular pollutants can be found in Appendix C.
c) I is the Stacked Filter Unit (Cahill) and C, L, and S are the Automated
Dichotomous Samplers (Loo).
Column 5 in Table 13 gi"es the mean ratio for three of the elements char-
acteristically found in the fine fraction, while column 9 gives the mean ratio
for three elements characteristic of the crustal fraction. It appears that
tandem filters M and N obtain 11% more fine pollutant concentration than D and
R, the manual dichotomous samplers while the stacked filter unit obtains 13%
43
-------
less than the automated dichotomous sampler. The small percentage differences
for Zn, Pb, and S are not the focus of the particle bounce question. Continu-
ing, column 9 shows that M and N record twice as much Si, Ca, and Fe as do
samplers D and R, while sampler I records almost twice as much as samplers
C, L, or S. Finally, if the mean coarse results are ratioed to the mean fine
results (last column), any sampler bias or normalization differences are re-
moved. The average mean ratio shown in the bottom section of Table 13 indi-
cates that there is about twice as much coarse particle content in the fine
fraction for the tandem and stacked filter units relative to the manual and
automated dichotomous samplers. Once again, the very excellent agreement in
pollutant concentrations obtained by all of the sampler types at the study for
fine lead, sulfur, and zinc would indicate that the crustal particle increases
in the fine fraction observed for the tandem and stacked filter units can be
attributed to particle bounce from the coarse filter onto the fine filter.
There is also some evidence for particle bounce in another sampler that
collects particles by inertial impaction. The CHAMP sampler uses a 26 ^m cut-
off inlet, a glass fiber filter covered impaction stage and backup filters
designed to collect particles in the 0-3.5 ^m and 3.5-26/^m ranges. Compared
with other samplers the CHAMP sampler collected more mass in the fine fraction
and relatively less mass in the coarse fraction. This is consistent with a
particle bounce phenomenon that has been observed in previous field studies
when dry surfaces were used as impaction plates (12). Because of the bounce
problem, overestimation of the crustal materials in the fine fraction can
occur depending on the sampler type, and any attempt to measure the acidity
of the fine particle fraction would be compromised due to the presence of
alkaline crustal elements.
LOCAL SOURCES OF COPPER
Thirty nine samplers were assembled on the roof of the Federal Building
within approximately 18 meters. This is an average of one-half meter separa-
tion between samplers, much closer together than samplers would normally
operate in monitoring programs. Precautions were taken to vent the exhaust of
selected samplers over the edge of the roof if they were suspected of causing
problems. Several participants claimed filtered exhausts. In spite of the
precautions taken there appears to have been several localized sources of cop-
per, probably from winding dust generated in several sampler pump motors.
Copper and zinc are neighboring elements in the periodic chart (Z = 29
and 30, respectively) and their mean concentrations throughout the^study were
comparable (90+40 and 120+20 ng/m - totals, and about 80 ng/m each-
small). Their respective K x-rays are not considered soft and are easy to
excite via the several x-ray fluorescence analysis techniques used to obtain
the concentrations from the collected particulate matter. Thus, they would
appear to be excellent monitors of each other if one assumes their behavior
as aerosols are similar. As has already been mentioned in the previous sec-
tions, the zinc pollutant levels were, along with sulfur and lead, among the
most consistently and accurately measured. This is illustrated by the two
graphs shown in the top half of Figure 10. These plots show the 16 period
average ratios for total and small zinc obtained by all those samplers re-
porting trace elements. The standard deviations on the total fraction are
44
-------
T3
O CL)
• i —
IIII
- N t
Ei f\
! O
CO
1 C
H
hO
h-0— I
IIII
IIII
-\ |
i —
h-0— | -
1
1 1 1
N «- U
C
IIII
— c
N
1-
1 r
i y-\
- h-O-
H>
L /-\
IIII
IIII
•\ I
(-OH
i
h-CM -
H —
i
o
CO
_J
o
cc
z
Q
•> c
3
—>
I
O
CO
_l
o h
QC
o»i_
N >— LO O
d
SOUBJ aBej;
IIII
_ D
ro
E
CO
1
t
I
I
IIII
—
-0—1 -
IIII
M «— u
C
IIII
o
"S .
— -^ 1
i
IIII
^
IIII
03-1-03
O) -E U
— CLl— O
CL i—
03
O • "O
E E 00
-13 +J U
, 03 03
, „, S- =3 -E
1 >^ •+-> -t-> O
^ E 0 -r-
, -. Q. CD 13 -E
U E 0 r- S
CD C f-
_ CO O 00
"- O CD •+->
en oo
z 0 S- Z5
, E O3 03
.J K D M ^
" zinc and copper. The zi
jper concentrations show
jme sampler pump motor exh
, C, and S.
0 0 i— Q
<+- O O
— > 00
oo D i s-
O i — .E <1>
I -r- -i- CTli —
1 -4-> -E -I- Q.
03 S -C E
,n S- 03
1 "V ">
1 S- • - J2'
CU T3 S- S-
CO fe i— OJ 03 O
— Q. E
to a) -»-> o
OJ -M D.
cj en cu a) co
03 -O ^
S- T3 -
(T OJ >> -
03 O) i— O
2 "O 03 03 -
O S- J3 -
«_ -i- = 0
5 S- O S- S-
— i -Q S O
J - ^ d
O r-H
3AV OJ
CD
45
-------
comparable to those for the small fraction (indicating that even the large
zinc particles were well behaved).
These two graphs are to be compared to the lower half of Figure 10 which
shows similar results for copper. Note that samplers M, N, and L report ap-
proximately equivalent concentrations for total and small, zinc and copper.
These samplers are centrally located. Samplers C and S report twice as much
small and large copper as L, while the nearly adjacent samplers D and R report
less than twice as much as M and N. This decrease in recorded concentration
towards the center suggests the outer samplers A and/or B, and T and/or U -
all CHAMP or high volume samplers - may be responsible for the localized cop-
per source. These samplers have the highest flow rates. Two additional high
volume samplers were in permanent operation by the West Virginia Air Pollution
Board about 2 meters to the southwest of sampler U. This makes a total of 3
high volume samplers to the southwest of sampler S, and only one to the north-
east of sampler C, yet the concentrations recorded by S and C are nearly the
same. This suggests a rapid decrease in localized concentrations levels with
distance. Note also that sampler D was 0.7 m inwards and adjacent to C, while
R was 1.3 m inwards and NOT adjacent to S (see Table 3). Since no trace ele-
ment concentration levels were evaluated for any of the CHAMP or high volume
filter samples, it is not known whether these samplers can induct their own
emissions. This can probably happen, but will be dependent on meteorological
conditions.
There also appeared to be a strong local source of copper and its concen-
trations were highest during some of the even periods (2000-0800 hrs). From
the diurnal wind patterns this places the source southeast of the Federal
Building. For periods 2, 4, 12, 14, and 16 mean copper concentrations varied
from 80 to 400 ng/m and averaged 200 ng/m . The other three even periods av-
eraged 38+6 ng/m ; while all odd periods except 13 and 15 averaged 28+6
ng/m (13 and 15 averaged 54+6 ng/m ). Eliminating all results from the
five high concentration even periods, and calculating an average concentration
value for samplers (D, C, and S) less samplers (M, N, R, L, G, H, J, and K)
yields 32 ng/m additional total copper collected by samplers D, C, and S.
Almost 95% of the total copper seen during periods 2, 4, 12, 14, and 16 was
contained in the small fraction (^ 3.5^m); whereas only about 60% of the pump
motor copper is contained in the small fraction.
BROMINE LOSS
Thirty fine particulate samples were selected from one of the automated
dichotomous samples. These particulate samples were collected on teflon fil-
ter substrates having 1 //m pore sizes and were used as analytical technique
quality control samples. Results obtained by the participants were presented
above. It was mentioned that a possible bromine loss with time may have oc-
curred on these substrates. This evidence comes from four measurements.
First, the original measurement of bromine concentration on all 30 samples by
Loo when normalized to the lead concentrations gave 1.19 + 0.10. Second, the
LLL value obtained for Br to Pb one month later gave 1.14 + 0.03. Two months
later 5 of the 30 samples when remeasured after cutting gave 1.02 + 0.10 for
the Br to Pb ratio. Finally, six months after Loo's original measurement, he
remeasured 2 of the 24 samples that were mailed out and obtained 1.00 + 0.03
46
-------
for the relative bromine to lead concentration. In this last measurement Loo
also observed a 12% loss of chlorine. Thus, it appears that fine fraction
lead bromochloride deposited on teflon filter substrates may require two to
six months to stabilize after deposition.
DIURNAL VARIATIONS IN POLLUTANT CONCENTRATIONS
Because of the regular and dramatic wind shifts of nearly 180° at ap-
proximately 9 am and 9 pm each day, it is worthwhile examining the reported
pollutant concentrations for any day/night variations in the size fraction
concentrations. These variations are easily obtained from the composite sam-
pler concentrations or Z-means for the totals and two size fractions reported.
Odd periods correspond to daylight hour collections (0800-2000 hrs); while
even periods correspond to predominantly nighttime collections (2000-0800
hrs).
Table 14 lists the percentage of selected pollutants found in either the
small or large fraction. The tabular entries are recorded in percent. The
second column lists the 16-period average percent of pollutant concentration
measured in the small fraction, exceptions as footnoted. The results under
column three indicate that perhaps as much as 12% more small fraction mass is
collected at night as during the day. The crustal elements are definitely
concentrated in the large fraction, and only a suggestion of any diurnal
variation is found. Note that the sum of the small and large fractions for
the silicon and titanium percentages do not total to 100%. This bias is
introduced by samplers which report only total trace elements. The local
sources of copper were discussed earlier. Selenium is not entered in this
table. Its concentrations were very low, were concentrated 88% in the small
fraction, and showed no diurnal variation. During periods 13 and 11 selenium
concentrations increased to 2 and 3-times the typical concentrations of
lOng/m observed. Zinc also showed three excursions above normal during peri-
ods 2, 4, and 8. Otherwise about 60% of it is found in the small fraction.
Bromine and lead are two well known pollutants associated with the internal
combustion engine. They show evidence of a diurnal variation. The recorded
bromine concentration did increase to about twice the average observed con-
centrations during period 2, indicating a local non-automotive source. There
appears to be 20% more bromine collected at night. One possible explanation
is that the lack of sunlight allows more bromine to remain in particulate
form, while during the day more of it is converted to a gaseous form.
Period 8 and 9 appear to yield anomalous Z-mean ratios for all of the
crustal elements. These two periods were also the least normal in terms of
the typical diurnal wind patterns (see Figures 4 and 5); and period 9 was
unusual because the wind was nearly calm and shifted constantly (see Fig. 5).
47
-------
TABLE 14. PERCENT OF POLLUTANT IN THE SMALL OR LARGE FRACTION
Pollutant
Mass
S + S04
Silicon
Calcium
Titanium
Iron
Copper
Zinc
Bromine
Lead
Br/Pb
% in Small
Fraction
51 + 7b>
90 + 3
11 + 3b)
12 + 2b)
17 + 2b)
20 + 3b)
% in Small or
Large Fraction ,
Day, Night, Fraction ' Comments
48 + 6, 54 + 8, smallb) Pd. 8/9 high, 67% small
No diurnal effect Pd. 8/9 normal
ab)
76 + 5, 84 + 3, large'
87 + 2, 93 + 4, largeb)
86 + 6, 89 + 2, largeb^
77 + 3, 84 + 3, largeb)
Pd. 8/9 low, 57%-large
Pd. 9 low, 68%-large
Pd. 8/9 low, 72%-large
Pd. 8/9 low, 68%-large
See discussion on local sources of copper, page 44
57 + 7c) 55 + 7, 60 + 7, small Pd. 8/9 normal
79 + 9d) 72 + 7, 86 + 3, small 20% more Br at night
77+4 74 + 2, 79 + 3, small Pd. 8/9 normal
20 + 4d) 18 + 3, 23 + 3, small
a) The plus or minus value is the standard deviation. Small or Large refers
to the percentage of either the small or large fraction (Z-mean) found
relative to the total (Z-mean).
b) Periods (Pd.) 8 and 9 were omitted from calculations because of anomalous
percentages and unusual meteorology (see text).
c) Periods 2, 4, and 8 were omitted from the calculation. During these peri-
ods 87, 79 and 79% respectively, of the total zinc was recorded in the
small fraction. A local source of zinc probably caused the dramatic in-
crease in zinc concentrations recorded during these periods.
d) Periods 2 and 10 were omitted from the calculation. The bromine concen-
tration increased dramatically during period 2. The bromine to lead
ratios went to 61 and 33%, respectively, for periods 2 and 10.
48
-------
REFERENCES
1. Cooper, J. A. Workshop on X-Ray Fluorescence Analysis of Aerosols.
Battelle Northwest Laboratories Report, BNWL-SA-4690, 1973.
2. Proceedings of this conference published in: X-Ray Fluorescence Analysis
of Environmental Samples. T. G. Dzubay, ed., Ann Arbor
Science, Ann Arbor, Mich. 48106, 1977, 310 pp.
3. Kanawha Valley Air Pollution Study, EPA-APTD 70-1, U.S. Environmental
Protection Agency, Research Triangle Park, N.C. 27711, March 1970,
380 pp.
4. Dixon, W. J. Processing Data for Outliers. Biometrics: 22, 74-89,
1953.
5. Tietjen, G. L. and R. H. Moore. Some Grubbs-Type Statistics for the
Detection of Several Outliers. Technometrics, 14: 583-597, 1972.
6. Camp, D. C., J. A. Cooper, and J. R. Rhodes. X-Ray Fluroescence
Analysis - Results of a First Round Intercomparison Study. X-Ray
Spectrometry, 3: 47-50, 1974.
7. Camp, D. C., A. L. Van Lehn, J. R. Rhodes, and A. H. Pradzynski.
Intercomparison of Trace Element Determinations in Simulated and Real
Air Particulate Samples. X-Ray Spectrometry, 4: 123-137, 1975.
8. Stevens, R. K., T. G. Dzubay, G. Russwurm, and D. Rickel. Sampling and
Analysis of Atmospheric Sufates and Related Species. Atmospheric Envi-
ronment (in press), 1978.
9. Contant, R. W. Factors Affecting the Collection Efficiency of Atmos-
pheric Sulfate. U.S. Environmental Protection Agency, Office of Research
and Development Report EPA-600/2-77-076, Research Triangle Park, NC
1977.
10. Spicer, C. W., P. M. Schumacher, J. A. Kouyoumjian, and D. W. Joseph.
Sampling and Analytical Methodology for Atmospheric Nitrate. U.S.
Environmental Protection Agency, Office of Research and Development
Report EPA-600/2-78-009, Research Triangle Park, 1978.
11. John, W., G. Reischl, S. Goren, and D. Plotkin. Anomalous Filtration of
Solid Particles by Nuclepore Filters. Atmos. Environ. (To be published).
12. Dzubay, T. G., L. E. Nines, and R. K. Stevens. Particle Bounce Errors in
Cascade Impactors. Atmos. Environ. 10: 229-234, 1976.
49
-------
R. Burton
T. Cahill
R. Delumea
T. G. Dzubay
G. M. Hudson
B. W. Loo
P. Mueller
C. E. Rodes
R. L. Tanner
R. Burton
r. Cahill
R. Delumea
T. G. Dzubay
G. M. Hudson
B. W. Loo
P. Mueller
C. E. Rodes
R. L. Tanner
APPENDIX A
SAMPLER AND ANALYTICAL TECHNIQUE SUMMARIES
PART I
One Page Summaries
CHAMP and High Volume Samplers 52
Sierra Multiday and Stacked Filter Units 54
TWO MASS sequential tape sampler 57
Manual Dichotomous and Series Filter Samplers ... 58
Linear Streaker Samplers . 60
Automated Dichotomous Samplers 61
Cyclone Sequential Filter Samplers 62
EMSL Dichotomous and High Volume Sampler 63
Diffusion and High Volume Sample 65
PART II
Descriptive Summaries
CHAMP and High Volume Samplers 67
Sierra MuKiday Stacked Filter Units ........ 69
TWO MASS sequential tape sampler 74
Manual Dichotomous and Series Filter Samplers ... 76
Linear Streaker Samplers 81
Automated Dichotomous Samplers 84
Cyclone Sequential Filter Samplers 86
EMSL Dichotomous and High Volume Sampler 90
Diffusion and High Volume Sample 92
51
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Robert M. Burton
INSTITUTION: U.S. EPA, Health Effects Research Laboratory
SAMPLER TYPE: CHAMP Participate Fractionator
OPERATING CHARACTERISTICS:
FLOW RATE: 1130 t/m
PARTICLE SIZE CUTPOINTS: 26 pm (upper cut-off)
3.5pm (lower cut-off)
COLLECTION SUBSTRATE: 12-inch diameter Glass Fiber,
Gel man Type A (3.5-26pm)
8-inch by 10-inch Glass Fiber,
Gelman Type A (0-3.5 pm)
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
1. Sierra Calibration Orifice Model 330.
2. Sierra High Volume constant Flow Controller model 310, with measured
stability of .03 cubic meters/minute.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
1. Mass (0-3.5pm).
2. Mass (3.5-26pm).
3. Suspended Nitrates (0-3.5pm).
4. Suspended Sul fates (0-3.5pm).
ANALYTICAL TECHNIQUE(S) USED:
1. Mass - (0-3.5pm) and (3.5-26pm) Mettler Digital Balance (.01 mg
sensitivity).
2. Suspended Nitrates - (0-3.5pm) Filter strip reflux with distilled
water, copper-cadmium reduction, sulfanil amide reaction.
3. Suspended Sulfates - (0-3.5pm) Filter strip reflux with distalled
water, water soluble sulfate then measured by the methyl thymol blue
(MTB) method.
52
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Robert M. Burton
INSTITUTION: U.S. EPA, Health Effects Research Laboratory
SAMPLER TYPE: CHAMP Flow Controlled High Volume Sampler
OPERATING CHARACTERISTICS:
FLOW RATE: 1130 i/m
PARTICLE SIZE CUTPOINTS: none
COLLECTION SUBSTRATE: 8-inch by 10-inch Glass Fiber,
Gel man Type A
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
1. Sierra Calibration Orifice, Model 330.
2. Sierra High Volume Constant Flow Controller Model 310, with measured
stability of .03 cubic meters/minute.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Total Mass
2. Total Suspended Nitrates
3. Total Suspended Sulfates
ANALYTICAL TECHNIQUE(S) USED:
1. Mass - Mettler Digital Balance (.01 mg sensitivity).
2. Suspended Nitrates - Filter strip reflux with distilled water,
copper-cadmium reduction, sulfanilamide reaction.
3. Suspended Sulfates - Filter strip reflux with distilled water,
water soluble sulfate then measured by the methyl thymol blue (MTB)
method.
53
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Thomas A. Cahill, + Air Quality Group, Crocker Nuclear Lab.
INSTITUTION: University of California, Davis 95616
SAMPLER TYPE: Sierra Multiday Impactor
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/in): 24 liters/min
PARTICLE SIZE CUTPOINTS: Stage 1 - 4.3 microns - less than 20 microns*
Stage 2 - 0.78 microns - 4.3 microns
Stage 3 - 0.01 microns - 0.78 microns
COLLECTION SUBSTRATE: Mylar type S, with about 60 micrograms/cm2
Apiezon L coatings; filter - 0.4 micron
Nuclepore
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Rotometer in instrument; orifice (0.5" H20) on
intake, (interm.) cal. by 9 litre Collins
Spriometer
No flow control (early unit borrowed for test)
Mean change in flow - 5.7%; Estim. error, + 2%
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Routinely included in output: Na,Mg,Al,Si,P,S,C1 ,K,Ca,Ti,V,Cr,Mn,Fe
Co,Ni,Cu,Zn,Ge,Ga,As,Se,Br,Rb,Sr,Zr,Mo,Ba,Pt,Au,Hg,Pb
2. All other elements heavier than sodium - x-ray lines recorded, with
energy and intensity, for manual reduction. Second analysis with
higher gain to cover Ca - Rare earths.
ANALYTICAL TECHNIQUES) USED:
1. PIXE - 18 MeV alphas (2 detector gains)
2. XRF - secondary fluorescers and filters
*Intake calibrated for 20 micron cut point (approximate); 60 mesh stainless
steel screen added since that time has reduced the cut point.
54
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Thomas A. Cahill, + Air Quality Group, Crocker Nuclear Lab.
INSTITUTION: University of California, Davis 95616
SAMPLER TYPE: Stacked Filter Unit
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 5 liters/min
PARTICLE SIZE CUTPOINTS: Coarse stage - 2.6 microns - less than 20
microns*
Fine stage - 0.01 microns - 2.6 microns
COLLECTION SUBSTRATE: Coarse stage - 8 micron Nuclepore Filter
Fine stage - 0.4 micron Nuclepore Filter
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Orifice meter intermittantly placed over
intake, calibrated by 9 litre Collins
Spirometer
Passive flow control through a 10:1 ballasting
orifice at pump.
Mean change in flow - 9%; Estim. error, + 2%
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Routinely included in output: Na,Mg,Al,Si,P,S,C1,K,Ca,Ti,V,Cr,Mn,Fe,
Co,Ni,Cu,Zn,Ge,Ga,As,Se,Br,Rb,Sr,Zr,Mo,Ba,Pt,Au,Hg,Pb
2. All other elements heavier than sodium - x-ray lines recorded, with
energy and intensity, for manual reduction. Second analysis done
during this test for Ca through rare earths.
3. Mass, coarse and fine
ANALYTICAL TECHNIQUE(S) USED:
1. PIXE - 18 MeV alphas (2 detector gains)
2. XRF - secondary fluorescers and filters
3. Mettler blance
*Intake calibrated for 20 micron cut point (approximate); 60 mesh stainless
steel screen added since that time has reduced the cut point.
55
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Thomas A. Cahill, + Air Quality Group, Crocker Nuclear Lab.
INSTITUTION: University of California, Davis 95616
SAMPLER TYPE: Total Filter Unit
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 24 liters/min
PARTICLE SIZE CUTPOINTS: None; intake, less than 20 microns*
COLLECTION SUBSTRATE: Gelman GA4 Filters (NOTE: Results obtained at
Charleston low by 8% due to incorrect filter
area used.)
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Orifice meter imtermittantly placed over
intake, directly calibrated by a 9 litre
Collins Spirometer
Passive flow control through a 2:1 ballasting
orifice at pump.
Mean change in flow - 1%; Estim. error, + 1%
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Routinely included in output: Na,Mg,Al ,Si,P,S,C1 ,K,Ca,Ti,V,Cr,Mn,Fe,
Co,Ni,Cu,Zn,Ge,Ga,As,Se,Br,Rb,Sr,Zr,Mo,Ba,Pt,Au,Hg,Pb
2. All other elements heavier than sodium - x-ray lines recorded, with
energy and intensity, for manual reduction. Second analysis done
during this test for Ca through rare earths.
3. Mass
ANALYTICAL TECHNIQUE(S) USED:
1. PIXE - 18 MeV alphas (2 detector gains)
2. XRF - secondary fluorescers and filters
3. Mettler blance
*Intake calibrated for 20 micron cut point (approximate); 60 mesh stainless
steel screen added since that time has reduced the cut point.
56
-------
SUMMARY OF SAMPLING AND ANALYSIS METHOD
PARTICIPANT: Richard Delumyea & Ed Macias
INSTITUTION: Washington University
SAMPLER TYPE: TWOMASS two stage sequential tape sampler
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 12 to 18 i/m
PARTICLE SIZE CUTPOINTS: 3.5^m
COLLECTION SUBSTRATE: Glass fiber with celluose backing
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
See long write-up, page 74
SPECIFIC POLLUTANT SPECIES MEASURED:
Mass, Sulfur
57
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: T. Dzubay, et al
INSTITUTION: US EPA
SAMPLER TYPE: Two manual Dichotomous, Samplers D and R
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 14 t/m
PARTICLE SIZE CUTPOINTS: 3. 5 ^m (14 urn upper cut-off)
COLLECTION SUBSTRATE: Sampler D: asymmetric teflon (1-10
Sampler R: 1 //m (FALP) Fluoropore
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Sampler D: Differential pressure regulator on pump exhaust
Sampler R: Sierra Series 250 servo system
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
Sampler D: mass and all elements detected by XRF
Sampler R: mass, $04, NO?, S, Pb, and other trace elements detected
by XRF
ANALYTICAL TECHNIQUE(S) USED:
Gravimetric, XRF, Ion chromatograph
58
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: T. Dzubay, et al
INSTITUTION: US EPA
SAMPLER TYPE: Two Tandem Filter Samplers, M and N
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 7.2 i/m
PARTICLE SIZE CUTPOINTS: 3.5
COLLECTION SUBSTRATE: Sampler M: 9.6 /im Nucleopore (coarse)/
0.4 fim Nucleopore (fine
Sampler N: 9.6 /;m Nucleopore (coarse)/
Assymetric Teflon (fine
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Sierra flow servo system
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
All elements detected by XRF
ANALYTICAL TECHNIQUE(S) USED:
XRF
59
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: G. Martin Hudson, Physics Department
Alistair C. D. Leslie, Oceanography Department
INSTITUTION: Florida State University
SAMPLER TYPE: Linear Streaker Samplers
OPERATING CHARACTERISTICS:
Streaker #1 Streaker #2
FLOW RATE (in liters/m): 0.5 i/m 0.35 l/m
PARTICLE SIZE CUTPOINTS: Total filter Total filter
COLLECTION SUBSTRATE: 0.4 p nuclepore 0.2 fj nuclepore
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Calibrated TYLAN model FC 260 Mass Flow Controllers. Field checks made
with rotometers.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
Particulate, elemental: Al, Si, S, Cl, K, Ca, Sc, Ti, V, Cr, Mn, Fe,
Co, Ni, Cu, Zn, Ge, As, Se, Br, Rb, Sr, Mo, Ag, Cd, Sn, Sb, I, Cs, Ce, Au,
Hg, Pb, Bi
ANALYTICAL TECHNIQUE(S) USED:
PIXE (proton induce x-ray emission) calibrated to commercially prepared
standards good to + 5%.
60
-------
SUMMARY OF SAMPLING AND ANALYSIS METHOD
PARTICIPANT: Billy W. Loo
INSTITUTION: Lawrence Berkeley Lab, Univ. of California
SAMPLER TYPE: 3 Automated Dichotomous Air Samplers: C, L, and S
OPERATING CHARACTERISTICS:
SAMPLING FLOW RATE: 50 f/m
PARTICLE SIZE CUTPOINTS: 2.4/^m
COLLECTION SUBSTRATE: O.l^m pore size teflon membrane filter
mounted on 5 x 5 cm plastic frames
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
The Dwyer RMC-103-SSV and Metheson #603 rotometers were used for
field calibration. They were cross-checked with laboratory dry and
wet test meters.
An automated flow controller consisting of pressure sensors and water
driven micrometer valves maintains the flow to within 1% of present
values.
SPECIFIC POLLUTANT SPECIES MEASURED:
Total mass, Si, S, Ca, Ti, Fe
Cu, Zn, Se, Br, and Pb
ANALYTICAL TECHNIQUE(S) USED:
Beta gauge, Energy dispersion x-ray fluorescence analysis, secondary
exciters.
61
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Peter K. Mueller
INSTITUTION: Environmental Research & Technology, Inc.
SAMPLER TYPE: Sequential Filter Samplers: E and F
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 100 t/m
PARTICLE SIZE CUTPOINTS: Sampler E: About 15 m at inlet
Sampler F: 2.5 /^m
COLLECTION SUBSTRATE: 47 mm dia. Teflon-coated glass fiber,
(Pallfelx TX40HI20)
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Calibrated orifice used for field calibration.
Flow monitored by continuous chart recording of vacuum upstream pump.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
Total mass, SO^, N03
Small mass, S0d, NO?
T1 O
ANALYTICAL TECHNIQUE(S) USED:
Mass: Gravimetry
SO^, N03: Ion chromatography
62
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Charles E. Rodes
INSTITUTION: EPA/EMSL
SAMPLER TYPE: EMSL Dichotomous
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 14 i/min
PARTICLE SIZE CUTPOINTS: Inlet 20/jm, outpoint 3.
COLLECTION SUBSTRATE: Fluropore 37 mm
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Flow Calibration using mass flow meters.
Flow maintained by restrictor valves.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
TSP, Pb, S04=
ANALYTICAL TECHNIQUE(S) USED:
TSP gravimetric
Pb Atomic Absorption (flameless)
S0^~ Ion Chromatograph
63
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Charles E. Rodes
INSTITUTION: EPA/EMSL
SAMPLER TYPE: Hi-Vol
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): 1415 t/m
PARTICLE SIZE CUTPOINTS: "Total Suspended Participates" (TSP)
COLLECTION SUBSTRATE: Glass Fiber 8 x 10 inch
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Flow Calibration using orifice meter.
Flow maintained by mass flow controller.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
TSP, Pb, S04=
ANALYTICAL TECHNIQUES ) USED:
TSP: gravimetric
Pb: Atomic Absorption
S04=: Methyl Thymol Blue
64
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Roger L. Tanner, Leonard Newman
INSTITUTION: Brookhaven National Laboratory
SAMPLER TYPE: Diffusion Sampler
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/m): DBO: 10.6 i/min; DB1 : 16.8 f/m
DB2: 13.8 i/m1n.
PARTICLE SIZE CUTPOINTS: DBO: 0 to ca 5 ^m; DB1 : 0.03 pm to ca 5/jm;
DB2: O.lZim to ca 5
COLLECTION SUBSTRATE: Phosphoric acid- treated Pall flex GAO tissue
quartz.
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Flows were calibrated with open Matheson rotometer tubes in series with
and upstream from the 47 mm filter holders. The flow variations in calibra-
tions before and after the Intercomparison averaged 11% for the 3 sample sets.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Sulfate
2. Nitrate
3. Strong acid
4. Ammonium
5. Sulfuric acid
ANALYTICAL TECHNIQUE(S) USED:
1. Extraction into pH 4 aqueous solution and determination by ion
chroma togrphy.
2. Hydrazine reduction and colorimetry.
3. Gran titration with correction for pH 4 leach solution.
4. Indophenol colorimetry.
5. Benzaldehyde extraction and determination by flash volatilization-FPD
or by ion chroma tography.
65
-------
SAMPLING AND ANALYSIS METHOD SUMMARY
PARTICIPANT: Roger L. Tanner, Leonard Newman
INSTITUTION: Brookhaven National Laboratory
SAMPLER TYPE: Staplex HiVol with 5-in circular filter pack in conventional
HiVol enclosure.
OPERATING CHARACTERISTICS:
FLOW RATE (in liters/in): 700 t/m
PARTICLE SIZE CUTPOINTS: 0 to ca 20 fim
COLLECTION SUBSTRATE: Phosphoric acid-treated Pall flex GAO tissue
quartz for aerosol collection. ^CO^, glyc-
erol-impregnated cellulose (S & S 2W, 2
sheets) for S02 collection.
METHOD OF FLOW CALIBRATION AND MAINTENANCE:
Magnetohelic gauge measured with filter pack in place with unexposed
filters; checked for repeatability with representative exposed filter_packs
after study. The average reduction in flow during sampling was 1.1 m /hr
( 3%) with average flow used for concentration calculations.
ALL SPECIFIC POLLUTANT SPECIES MEASURED:
(including those not intercompared)
1. Sulfur dioxide
2. Sulfate
3. Nitrate
4. Strong acid
5. Ammonium
6. Sulfuric acid
ANALYTICAL TECHNIQUE(S) USED:
1. Filter pack, extraction into peroxide solution, determination as sul-
fate by trubidimetry or ion chromatography.
1. Extraction into pH 4 aqueous solution and determination by ion
chromatogrphy, or extraction into 10" N HC1 with determination by
Methyl thymol Blue colorimetry.
2. Hydrazine reduction and colorimetry.
3. Gran titration with correction for pH 4 leach solution.
4. Indophenol colorimetry.
5. Benzaldehyde extraction and determination by flash volatilization-FPD
or ion chromatography.
66
-------
CHAMP AND HIGH VOLUME SAMPLERS
Robert Burton
Health Effects Research Laboratory
Environmental Protection Agency
Research Triangle Park, NC 27711
EQUIPMENT DESCRIPTION
Two CHAMP RSP samplers and one flow-controlled high volume sampler were
sent to Charleston, W. Va., to participate in an intercornparison study of
particulate analyzers. The CHAMP RSP sampler divides the particulate matter
into two collected groups, 3.5 - 26 microns and 0 - 3.5 microns.
Particles greater than 26 microns in diameter are excluded by the aero-
dynamic construction of the analyzer. Particulate matter in the 3.5 to 26 mi-
cron region is impacted on a 12-inch diameter filter utilizing a single stage
Andersen impactor plate. Particles smaller than 3.5 microns are collected on
a standard 8 X 10 inch hi volume filter. The Andersen impactor is adapted to
fit between the after filter and the base section of the fractionator itself.
All filter media are of glass fiber construction.
The third sampler was a standard hi volume sampler, which collects
all particulate matter smaller than 100 microns. All samplers used Sierra
Flow controllers operating on General Metal's Model 2000H motor assemblies to
maintain a constant 1.13 nr/min (40 cfm) flow. The hi volume sampler is des-
cribed in detail in Volume 36, No. 228 of the Federal Register.
SAMPLER CALIBRATION
All devices were calibrated using a standard sierra orifice calibrated
at the factory and verified by our laboratory Rootsmeter. The calibration
curve of the orifice is: Q = 1.4054 (AP)1/^where AP is the pressure differ-
ence in inches of water as measured across the orifice. Q is then corrected
using the following equation to 760 mm Hg and 25°C (298°K):
Q- = (Pa/Ta)1/2 (3.156)Q where Q,
is the standard flow,
Pa is the barometric pressure in mm
HG, and
Ta is the ambient temperature in
degrees Kelvin.
The orifice was placed on top of the Andersen head without the RSP top
hat and a calibration curve was drawn using clean filters for the Andersen and
hi volume heads which are assembled in series. Various flows were obtained by
67
-------
changing the speed of the hi vol motor with a variac to simulate different
loadings on the filters. A curve was obtained for each sampler. For field
operations a small rotameter was attached to the bottom of each hi vol and
calibrated simultaneously with the orifice. The rotameters facilitated taking
field flow measurement, although both rotameter and orifice readings were re-
corded for this study. After the response equations were determined for each
device the flow controllers were then set to obtain 1.13 m /min (40 cfm) flow.
The flow obtained during the study show little variation from this 1.13 set-
ting, varying only from 1.17 to 1.11 m /min. Keep in mind, however, that only
stop and start flows were recorded during operation and actual flows during
sampling may vary but should have been kept within the above tolerances with
properly operating flow controllers.
ANALYSIS
The glass fiber filters were conditioned at 40% relative humidity and
25°C for 24 hours. They were then numbered and weighed to the nearest 0.1 mg.
Upon return from the field they were again equilibrated at the previous condi-
tions and weighed. Total suspended particulate (TSP) and respirable suspended
particulate (RSP) were then determined for the appropriate samplers. The
weight of the collected particulate matter was divided by.the total volume of
air sampled to obtain TSP and RSP concentrations in ^tg/m . The sample volumes
were calculated using the following equation:
V = -—~-*— T where Qi = initial flow - m /min
Qf = final flow - nr/min
T = total sampling-time - min
V = air volume - m
After the particulate concentrations were determined, th_e filters were
then sent to_Stewart Laboratories in Knoxville, Tenn. for SO^ and NOo analy-
sis. The SO^ analysis was performed using the methyl thymol blue (MTB) method
while the N03 analysis utilized a copper-cadmium reduction column and NEDA
dye. The concentrations were expressed as//g/m of each ion.
Our program does not routinely analyze for lead. For specific lead
samples, Tom Dzubay would perform our lead analysis. Since his section also
participated in this study, no lead analyses were performed on these filters.
Further detailed explanation of these methods can be found in EPA pub-
lication 600/1-76-011, Community Health Environmental Surveillance Studies Air
Pollution Monitoring Handbook: Manual Methods.
68
-------
SIERRA MULTIDAY AND STACKED FILTER UNITS
T. A. Cahill
Crocker Nuclear Laboratory and the Department of Physics
University of California, Davis, CA 95616
The Davis Air Quality Group, working in conjunction with the California
Air Resources Board, the Research Applied to the National Needs Division of
the National Science Foundation, and the U.S. Energy Research and Development
Agency, had developed and modified techniques for collecting and analyzing
atmosperic particulates. Monitoring of aerosols in California (1) by size
and elemental composition for the California Air Resources Board (CARB), began
in January, 1973, using a Lundgren-type rotating drum impactor (2), modified
for this program, and since marketed as the Environmental Research Corpora-
tion, and later, Sierra Instruments Corporation, Multiday Impactor. Quanti-
tative operation of this device was achieved through use of quasi-mono!ayer
depositions on coated surfaces in the two impact!on stages, and non-hygro-
scopic, surface-deposition filter media for the backup filter. Normal inte-
gration period was 24 hours. Approximately 5000 station-days of data in three
size fractions, 20 (jm to 3.6 pm, 3.6 pm to 0.65 pm, and 0.65 to 0.01 jjm, have
been accumulated in this program. The major aims of the program are in char-
acterization of aerosol components, the identification of aerosol sources,
study of transport, transformations, and sinks, identification of sources of
gaseous pollutants measured by the ARB monitoring network, and effects of
aerosols and gases upon visibility.
Elemental analysis of samples collected in this program occurs after
transport of samples to Davis, with contamination and loss effects summarized
in the Table. Analysis occurs primarily via particle induced x-ray emission
(PIXE) (3,4) with significant support through energy-dispersive x-ray fluores-
cence (XRF) and ion scattering analysis (ISA). Smaller numbers of analyses
are done using a scanning electron microscope and an electron microprobe with
wave-length-dispersive XRF capabilities. Total number of elemental values to
date in the monitoring program exceed 500,000. The system has participated in
numerous inter!aboratory comparisons (5), in addition to a massive (6000 anal-
yses/year) internal program of analytical validation.
The need to perform diurnal and spatial profiles that have some size in-
formation and full gravimetric and elemental analysis compatibility has led to
development of low cost, portable stacked filter units (SFU's) (6) following a
suggestion of K. R. Spurney et al.(7) These units are designed to deliver a
quasi-respirable separation of aerosols, with a 50% cut point and shape as
close as possible to that occurring between the nasal-pharangeal and tracheo-
pulmonary compartments of the human respiratory tract. Evaluations of this
69
-------
device are given in the SFU table for a two stage device designed to operate
at 5 i/min.
In an attempt to upgrade the information delivered by the large number
of Hi-Vol samplers operated by state and local agencies in California, the SFU
has been evaluated as a continuous monitoring device. The SFU would operate
on a 7 day integration period, avoiding the statistical problems inherent in
a 1 day in 6 Hi-Vol mode, while delivering respirable and non-respirable size
cuts in samples suitable for gravimetric and full elemental analysis. Lower
flows are used, 2 i/min, and on occasion, especially when visibility problems
exist, a third stage is added to mimic the cut points of the multiday.
For the sampler intercomparison study, a Multiday impactor and a stacked
filter unit will be used. The impactor will use 0.6 mg/cm mylar, coated
with 0.065 mg/cm Apiezon type L grease, as impaction surfaces, and 0.4 micron
Nuclepore filter for the afterfilter. The stacked filter unit will use 8 mi-
cron and 0.4 micron filters. Flow for both units will be set by a spirometer-
calibrated orifice meter on the intake. Analysis will be by both PIXE and
XRF, with elements aluninum and heavier quoted either as observed values or
upper limits of elements not observed. Calibration of the analysis systems
will be via gravimetric thin element standards, and total accuracy will be
within + 10% absolute when statistical precision is adequate.
REFERENCES
1. R. G. Flocchini, T. A. Cahill, D. J. Shadoan, S. J. Lange, R. A. Eldred,
P. J. Feeney, G. W. Wolfe, D. C. Simmeroth, J. K. Suder, Env. Sci. and
Techn. 10, 76 (1976).
2. D. A. Lundgren, J. Air Poll. Contr. Assoc. 1_7, 4 (1976).
3. S. A. E. Johansson and T. B. Johansson, Nucl. Instr. and Meth. 137, 473
(1976).
4. T. A. Cahill, in "New Uses of Ion Accelerators," J. Ziegler, Editor,
Plenum Press, New York (1975).
5. D. C. Camp, A. L. Van Lehn, J. R. Rhodes, A. H. Pradzynski, J. X-Ray
Spectr. 4, 123 (1975); PIXE, Group # 15; SRF, Group # 4.
6. T. A. Cahill, L. L. Ashbaugh, J. B. Barone, R. A. Eldred, P. J. Feeney,
R. G. Flocchini, C. Goodart, D. J. Shadoan, and G. W. Wolfe, "Analysis of
Respirable Fractions in Atmospheric Particulates via Sequential Filtra-
tion," J. Air Pol. Contr. Assoc. 27, 675 (1977).
7. K. R. Spurney, J. P. Lodge, Jr., E. R. Frank, and D. C. Sheesley, Env.
Sci. and Techn. 3, 464 (1969).
70
-------
Summary of Sample Collection Evaluations
MULTIDAY IMPACTOR
1. Large particle cut-off (50%)
(intake)
2. First stage cut-off (50%)
3. Second stage cut-off (50%)
4. Impaction collection efficiency
5. Filter capture efficiency
a) laboratory aerosols
b) ambient aerosols
6. Sizing errors due to non-
optimum ol lection
a) coarse collected on filter
(soils)
b) fine collected on coarse
(Pb, S aerosols)
7. Summary of sizing:
8. Flow calibration
9. Total unit efficiency
(fine particles)
Dp = 20 pm, variable w. wind speed*
Dp = 4.3 p, +0.4 urn
Dp = 0.78 pm, +0.05 urn
(97+3)%
(96+2)%
(99+3)%
Stage 1 (coarse) < 20 urn to 4.3 urn
Stage 2 (intermediate) 4.3 pm to
0.78 urn
Stage 3 (fine) 0.78 |um to 0.01 ym
Rotometer - continuous
Orifice and meter on Intake (0 to 1"
H20
(Intermittent; Spirometer calibrated)
(93+3)% (4 X 5a)
[correction not applied to the
Charleston data]
Summary of Sample Transport and Handling Corrections
1. Contamination in Handling
2. Loss of particles during
transport
^< 15 ng/m^ Si, _
<10 ng/nr Al, K; 6 ng/nr others
< 2% of Si value at most locations
less for other elements
< 2%
3. Loss of material during storage «10%
*Addition of a 60 mesh screen has lowered this cut.
71
-------
Summary of Sample Collection Evaluations
U.C. Davis Stacked Filter Unit at 5 1/min (2 stage)
1. Large particle cut-off (50%)
(intake)
2. Coarse stage cut-off (50%)
3. Filter collection efficiency
a) laboratory aerosols
b) ambient aerosols
4. Sizing errors due to non-
optimum collection
a) coarse particulates on fine
stage (soils)
b) fine particulates on coarse
stage (Pb, idling auto)
5. Summary of sizing:
6. Flow calibration
7. Total unit efficiency
D = 20 urn, variable w. wind speed*
D = 2.6 urn + 0.5 pm, quasi-respir-
able shape
(96+2)%
(99+3)%
< 6%
<2.5%
Coarse < 20 jjm to 2.6 urn
Fine 2.6 jum to 0.01 jum
Shape quasi-respirable, 20 1/min
nasopharynx
Orifice and meter on Intake (0 to 1"
H20
(Intermittent; Spirometer calibrated)
(96+2)% (4a)
[correction not applied to the
Charleston test]
Summary of Sample Transport and Handling Corrections
1. Contamination in Handling
2. Loss of particles during
transport
3. Loss of material during storage
<15 ng/m^ Si, ~
<10 ng/nr Al, K; < 6 ng/nr others
< 2% of Si value at most locations
less for other elements
«H
^Addition of a 60 mesh screen has lowered this cut.
72
-------
Summary of Analytical Corrections and Uncertainties
1.
2.
3.
4.
5.
6.
Statistical uncertainty in peak counts
Uncertainty in gravimetric standards
Integration of Ion beam
Ion beam attenuation
Electronic corrections, accuracy
Peak integration and background
uncertainties
Thus, reproducibility
system uncertainty
Loss of volatile elements (beam +
vacuum)
Particle Size Corrections
variable
t
+
5%
2%
5%
5%
+1 Q%
< 5%
(variable)
Stage #1
Stage #2
Filter Stage
Na 2.49 1.32
Mg 1.96 1.22
Al 1.68 1.16
Si 1.49 1.12
P 1.37 1.09
S 1.26 1.07
Cl 1.31
K 1.12
others <10% ~,~~
Loading Corrections (for 120 pg/nr aerosol)
Stage #1 Stage #2 Filter
Na <10% 1.70 1.22
Mg " 1.39 1.11
Al " 1.23 1.07
Si " 1.14 1.04
others
Stage
Uncertainty in particle size and loading corrections are estimated at + 2Q%
for most ambient aerosols, although, especially for the particle size correc-
tions, worse cases are common. These corrections thus increase the total
error for the four lightest elements.
REFERENCE
(Sierra Multiday)
"Monitoring California's Aerosols by Size and Elemental Composition",
R. G. Flocchini, T. A. Cahill, D. J. Shadoan, S. J. Lange, R. A. Eldred,
P. J. Feeney, G. W. Wolfe, D. C. Simmeroth, and J. K. Suder, Env. Sci. and
Technology K), 76 (1976).
REFERENCE
(Stacked Filter Units)
"Analysis of Respirable Fractions in Atmospheric Particulates via Sequential
Filtration", T. A. Cahill, L. L. Ashbaugh, J. B. Barone, R. A. Eldred,
P. J. Feeney, R. G. Flocchini, C. Goodart, D. J. Shadoan, and G. W. Wolfe,
J. Air Pollution Control Assoc. 27, 675 (1977)
73
-------
TWO MASS SAMPLER
R. Del Delumyea and Edward S. Macias
Department of Chemistry
Washington University
St. Louis, Missouri 63130
INTRODUCTION
As part of an EPA/DOE sponsored Sampler Intercomparison Field Study,
TWOMASS tape samplers were installed on the roof of the Federal Building in
downtown Charleston, West Virginia and run from May 11 through May 19, 1977.
One TWOMASS sampler collected 6 hr samples at a nominal flow rate of 12 ^ pm
and another collected 3 hr samples at a nominal flow rate of 18 ^pm. The 3
hr sampler was equipped with a beta-attenuation mass monitor system which pro-
vided a continuous record of coarse (>3.5 fjm) and fine (<3.5 urn) particle con-
centration.
The purpose of the study was to compare results obtained from various
sampling and analytical procedures. The sulfur content of the samples was
determined by flash vaporization followed by flame photometric detection
(FV-FPD). The sampling system was designed for high time resolution (30-120
min). For this study three hour samples were collected. Since the reporting
time for all intercomparison data was set at 12 hour intervals, four 3-hr sam-
ples were added to give the required information. In cases where one 3-hr
sample was not available, the remaining three were averages. If more than one
sample was missing during a twelve hour period, no results were reported.
ANALYTICAL METHODS AND PROCEDURES
Sampler
Atomospheric aerosol samples were collected in two size ranges with a
TWOMASS automated two-stage sequential filter sampler. The flow system sepa-
rated particles into two size fractions. Coarse particles (diam > 3 urn) were
impacted on a glass fiber filter with cellulose backing. Fine particles were
collected on the same type of high efficiency-low mass density filter. The
flow rate through the TWO MASS was set at 12 or 18 ^/min using a GAST rotary
vane pump.
Beta Attenuation Mass Monitor
Both the impaction and filtration heads of one TWOMASS were fitted with
independent sources (Carbon-14) and detector systems (solid state ruggedized
silicon surface-barrier type). After electronic amplification and filtration
74
-------
the output signals were sent through counters to a programmable calculator.
Mass was calculated using the equation
where AM is the mass concentration, A the spot area, f is the flow rate, ^im
the mass attenuation coefficient, t the time counting interval, IQ the count
rate of the previous interval and I the count rate of the current interval.
The beta intensity transmitted through the filter paper was measured as the
particulates were being deposited; mass concentrations were obtained at the
end of each 10-min counting period. Tapes were automatically advanced at
three hour intervals.
Particulate Sulfur Analysis
Water soluble particulate sulfur in the fine particle samples from the
TWOMASS instrument was determined using a flash vaporization-flame photometric
detection method. The sulfur analyzing system consisted of a flash vaporiza-
tion vessel, flame photometric detector (Meloy SA-160), electronic integrator,
and a strip chart recorded. The sample vaporization was performed by capaci-
tor discharge across a tungsten boat, resulting in resistance heating to
1100°C. Vaporized gaseous decomposition products of sulfur compounds were
carried to the flame photometric^detector by a stream of clean, charcoal fil-
tered air at a flow rate of 2 cm /sec. The linearized output was registered
on the strip chart recorder and the peak area integrated by an electronic
integrator.
To the filter deposits, 0.5 to 1 ml double distilled deionized water was
added and utrasonically disintegrated at 25°C for 10 minutes. Sample standard
solutions were transferred with a 5 \ii microsyringe to the tungsten boat and
heated at 60°C for 30 seconds until dry. The residue was then vaporized by
capacitor discharge.
This technique was calibrated using solutions of sulfuric acid, ammonium
sulfate and bisulfate, zinc- '"d zinc-ammonium sulfate in the range 0.16-34
ug/ml of sulfur. [Husar et al , Anal. Chem., 47, 2062 (1975)]. A 5 pi sample
of distilled deionized water gave a signal equivalent to 0.3 + 0.05 ng sulfur,
corresponding to a solution concentration of 0.06 pg/ml. This value was a
factor of five lower than the measured sulfur content of filter blank extract.
75
-------
MANUAL DICHOTOMOUS AND SERIES FILTER RESPIRABLE SAMPLERS
T. G. Dzubay, R. K. Stevens, and L. A. Nines
Atmospheric Instrumentation Branch
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
G. M. Russwurm and D. G. Rickel
Northrop Services, Inc.
Research Triangle Park, N.C. 27709
AEROSOL SAMPLING DEVICES
For our participation in the sampler intercomparison conducted in Charle-
ston, West Virginia between May 11 and May 19, 1977, four size fractionating
aerosol samplers were used. Two were manual dichotomous samplers (1) (MDS-S
and MDS-A), and two were tandem filter respirable samplers (2) (TF-M and
TF-P). Table 1 summarizes the characteristics of each sampler and the filter
media that were used.
The manual dichotomous samplers use a virtual impactor designed to sepa-
rately collect fine and coarse particles with a 50% separation diameter occur-
ring at 3.5 ym. The upper 50% cutoff occurs at 14 urn and is due partly to
losses in the virtual impactor but is mainly due to the use of a sampling in-
let designed to sample particles with an efficiency that is independent of
wind speed and also to reject giant, nonrespirable particles.(3)
The tandem filter sampler consists of a large pore sized Nuclepore filter
followed by a highly efficient filter. This sampler seprately collects the
coarse and fine particle fractions on the two filters.(2,4,5) For the flow
rates and filters shown in Table 1, the cutpoint occurs at 3.5 \m for liquid
particles.(2,4) Although there are no particle losses for this fractionator,
the size separation is not quantitative because of tendency for dry, solid
particles to bounce through the first filter.(6) The upper 50% cutoff of the
inlet is estimated to occur at about 20 pm.
Flow rates were checked using a Sierra Model 541* portable mass flow
meter at the beginning and end of each sampling interval. Such measurements
were made while the samplers were running and required about one minute per
*Mention of commercial products or company names does not constitute endorse-
ment by the U.S. Environmental Protection Agency.
76
-------
sampler. The calibration of the protable flow meter was checked daily against
a dry test meter at flow rates of 6 and 14 liters per minute. After the first
4 days of sampling the jets in the virtual impactor were inspected and cleaned.
CALCULATION OF CONCENTRATIONS
In determining the sampled volumes, the mean of the initial and final
flow rates was used. Samples were rejected if the final flow rates differed
by more than 15% from the nominal values indicated in Table 1. (See page 78.)
For each dichotomous sampler there is a small flow of air through the
coarse particle filter, and a proportionally small amount of fine particle
mass is collected on the coarse particle filter. To correct for this effect,
particle concentrations in the two size ranges were determined from the equa-
tions:
Cf = Mf/(tFf)
Cc = [Mc-(MfFc/Ff)]/[t(Ff + Fc)].
where C^ and Cc = the atmopheric concentrations of the fine and coarse par-
ticle fractions, respectively in jjg/m
Mf and MC = the masses collected on the fine and coarse particle fil-
ters, respectively in jug
Ff and FC = the flow rates through the fine and coarse particle fil-
ters, respectively in m/min
t = the sampling time in minutes.
MASS ANALYSIS
The masses of the aerosol deposits were determined gravimetrically using
a Perkin Elmer Ad-2 electrobalance that was located in the laboratory at
Research Triangle Park, N.C. The electrobalance was operated on the 100 mg
scale, and the weighing precision was 10 fjg. At the beginning of each daily
series of weighings, the span of the electrobalance was adjusted for correct
response to a standard 100 mg weight. The zero of the electrobalance was
checked before the weighing of each filter and was adjusted whenever a nonzero
reading occurred. To eliminate any electrostatic charges on the filters, they
were held for a few seconds within 1 cm of a Po radioactive source (500
juCi) before weighing. A second similar radioactive source was positioned
inside the balance enclosure. The relative humidity in the balance room was
between 35 and 45% for all weighings. To equilibrate the filters after sam-
pling, the petri dishes containing the filters were stored in the balance room
for a least 24 hours with the covers partly open.
ELEMENTAL ANALYSIS
Elemental analysis of samples by x-ray fluorescence spectroscopy was
carried out with an energy dispersive unit.(7) The detector is a Si(Li) diode
with a resolution of 208 eV for Mn-55 K x-rays. This spectrometer (a pulsed
x-ray tube design) excites the sample by characteristic x-rays from a selected
set of secondary fluorescers. Elements with atomic numbers 13 through 20 are
77
-------
TABLE A.I. SAMPLER AND FILTER MEDIA DESCRIPTIONS USED IN THE
INTERCOMPARISON STUDY.
Sampler
Type
Description
Upper 50% Flow Deposit
cutoff rate Area Filter Media
dia. 1/min. cm
Dichotomous
(MDS-S)
Dichotomous
(MDS-A)
Tandem Filter
(TF-M)
Tandem Filter
(TF-P)
Two stage virtual impactor 14jum 14
having 50% cut point at measured
3.5 pm. The flow rates
through the fine and coarse
filters is 13.6 and 0.4
liters/min respectively and
is maintained using a
Sierra Series 250 mass flow
sensor and pump servosystem.
Same as MDS-S except that
the constant flow rate is
maintained using a con-
stant differential pres-
sure controller on the
pump exhaust.
14 jum 14
measured
Two filters operated in
series with a sampling
velocity of 15 cm/sec.
maintained using a
Sierra (Kurtz) servo-
system.
Same as TF-M
20 jum 7.2
estimated
20pm 5.9
estimated
6.7 Teflon membrane
having ljum pores
bonded to a poly-
ethylene net
(FALP Fluoropore
from Millipore
Corp.)
6.7 Teflon membrane
having assyme-
tric pores (l^im/
10 ;jm) oriented
so that parti-
cles are col-
lected on the
1 jjm side
(P137PL25 Ghia
Corp.)
8.0 Coarse: Nucle-
pore having 9.6
jjm dia. pores
(N137PE2 Ghia
Corp.)
Fine: Nuclepore
having 0.4pi dia.
pores (N040 Poly
101 Ghia Corp.)
6.7 Coarse: 9.6 jum
Nuclepore (N137-
PE1, Ghia
Corp.)
Fine: Teflon,
same as used
for sampler
MDS-A above.
78
-------
analyzed using a titanium fluorescer; elements with atomic number between 22
and 38 plus Pb and Hg are analyzed using a molybenum fluorescer; elements Cd,
Sn, Sb and Ba are analyzed with a samarium fluorescer.
Spectral analysis was carried out with a linear least squares fitting
program which uses a library of single element spectra as the fitting func-
tions. These spectra were acquired by the spectrometer under analysis condi-
tions. A blank filter spectrum is used for the background and is included in
the library. (8) The spectrometer is calibrated for copper and sulfur using a
well characterized thin copper film standard and thin laboratory generated de-
posits of 0.3 jum copper sulfate particles. (8) Evaporated metal film standards
obtained from Micromatter Co. (Seattle, Washington) were used for lead and
other elements. (8)
To correct for attenuation of sulfur x-rays in the fine particle frac-
tion, a very slight correction (1 to 10%) was made in proportion to the amount
of mass collected. (8) For the coarse particle fraction, the attenuation cor-
rection factors for Si, S, K, Ti , Fe, Cu, In, Se, Br, and Pb, used were 0.48,
0.78, 0.87, 0.94, 0.94, 0.95, 0.98, 0.99, and 1.00 respectively. (9) Descrip-
tions of the analysis procedure are described in more detail by Stevens et
ANALYSIS OF IONIC SPECIES
The samples from MDS-S were analyzed for sulfate and nitrate using ion
exchange chromatography. Extraction, the initial step in the analysis was ac-
complished by placing each sample ig a Nalgene polypropylene bottle (30 ml
volume) containing 20 ml of 5 x 10" N perchloric acid extraction solution.
The extraction vessel was capped and put in a sonic bath for 20 minutes. A
study using x-ray fluorescence techniques has shown this procedure extracts
99% of the sulfur from the fine fraction and 95% of the sulfur from the coarse
fraction.
After extraction, sulfate and nitrate were determined using a Dionex
Model 14 Ion Chromatograph (1C). The samples were spiked with a base (.003M
NaHCOo + 0.0005M Na2C03) so that the range 10 M to 10~1/| was spanned. Reten-
tion times and peak areas were obtained using standards (10~ to 10 M of
(NH4)2S04 and NaNOo), and a calibration curve was drawn. Analysis of sulfate
data showed that the minimum detectable level was 10" neq/ml and that sulfate
concentrations were determined to within + 10%.
2
The minimum detectable level for nitrate was 5 x 10" neq/ml, and the
concentrations were determined to within + 155. For the 12 hour samples col-
lected with a 14 1/min flow rate, the detection Iimits3expressed in units of
atmospheric concentrations were 10" ug/m and 7 x 10 ug/m for sulfate and
nitrate, respectively. A more detailed description of the extraction and
analysis procedures is given by Stevens et al.(8)
79
-------
REFERENCES
1. Dzubay, T. G., Stevens, R. K., and Petersen, C. M. (1977), Application of
the dichotomous sampler to the characterization of ambient aerosols. X-
ray Fluroescence Analysis of Environmental Samples. Ann Arbor Sci. Pub.
Inc., Ann Arbor, MI 95-105.
2. Parker, R. D., Buzzard, G. H., Dzubay, T. G., and Bell, J. P. (1977), A
two stage respirable sampler using Nuclepore filters in series. Atmos.
Environ. 11, 617-621.
3. McFarland, A. R., Wedding, J. B., and Cermak, J. E. (1977), Wind tunnel
evaluation of a modified Andersen sampler and an all weather sampler in-
let. Atmos. Environ. 11, 535-539.
4. Parker, R. D., and Buzzard, G. H. (1978), A filtration model for large
pore nuclepore filters. Aerosol Science 9_ (to be published).
5. Cahill, T. A., Ashbough, L. L., Barone, J. B., Eldred, R. A., Feeney,
P. J., Flocchini, R. G., Goodart, C., Shadoan, D. J., Wolfe, G. W. (1977),
Analysis of respirable fractions in atmospheric particulates via sequen-
tial filtration. J. Air Poll. Cont. Assoc. 27_, 675-678.
6. John, W., Reischl, G., Goren, S., and Plotkin, D. (1978), Anomolous fil-
tration of solid particles by Nuclepore filters. Atmos. Environ, (to be
published).
7. Jaklevic, J. M., Loo, B. W., and Goulding, F. S. (1977), Photon induced
x-ray fluorescence analysis using energy dispersive detector and dichoto-
mous sampler. X-ray Fluorescence Analysis of Environmental Samples. Ann
Arbor Science, Ann Arbor, MI 3-18.
8. Stevens, R. K., Dzubay, T. G., Russwurm, G., and Rickel, D. (1978), Sam-
pling and analysis of atmospheric sulfates and related species. Atmos.
Environ, (to be published in January 1978).
9. Dzubay, T. G., and Nelson, R. 0. (1975), Self absorption corrections for
x-ray fluorescence analysis of aerosols. Advances in X-ray Analysis 18,
619-631.
80
-------
LINEAR STREAKER SAMPLERS
G. M. Hudson, A. C. D. Leslie, H. Kaufman and J. W. Nelson
PHYSICS AND OCEANOGRAPHY DEPARTMENTS
Florida State University
Tallahassee, FL 32306
LINEAR STREAKER SAMPLERS
The Florida State Elemental Analysis Group working with the Environmental
Protection Agency, the National Science Foundation, the Florida Sulfur Oxide
Study Board, and numerous international groups in Europe, Asia, Africa, and
South America has developed a complete air particulate monitoring system. The
heart of this system is the linear streaker.(1)
During the Charleston study four streakers were deployed, but data were
reported for only two of the samplers. These devices filter continuously col-
lecting particulate over a small area of 2 mm x 5 mm. An orifice is moved by
a synchronous motor-screw drive across the filter surface at a rate of 1 mm
per hour to produce a 5 mm wide, 168 mm long strip sample in a time of one
week. Analysis of the filter is limited to a width of 2 mm which yields pol-
lutant data equivalent to 84 time averaged 2 hour samples. For the purpose of
the intercomparison study 6 two hour samples were averaged to give the 12 hour
samples reported. Polycarbonate nuclepore filters of 0.4 pm and 0.2 jum pore
size were used depending on the flow rate desired. The Nuclepore filter it-
self may be used to limit the flow, or a flow control device may be connected
in series. Flows of one liter per minute or less produce satisfactory load-
ings for PIXE analysis of elements Z > 13. Regularly measured are Al, Si, S,
Cl, K, Ca, Fe, Zn, Br, and Pb; and less often Ti, V, Cu, Ni, Sn and As. The
streaker samples are also well suited for PESA(2) (proton elastic scattering
analysis) which gives trace element information in the range 1 < Z < 20.
The streaker is a simple reliable, and flexible sampler, which is also
highly portable and inexpensive. This has led to its reliable use in the Ama-
zon basin, in the Namib desert, and in the Indian Ocean. Over 100,000 elemen-
tal data points have been generated in one year with a time resolution needed
to compare with meteorological data. The linear streaker is a total filter
device. Its main disadvantage is that it does not give two size fractions.
This will be remedied by the circular streaker(3) which will soon be network
tested. The linear streaker's intake is vertically upward, and electron mi-
croscope pictures of filters indicate an upper size limit of 15 pm for parti-
cles collected. The Nuclepore filter is mounted on an aluminum frame which
can be quickly inserted and removed from the streaker.
81
-------
The streaker filters frames require no preparation before being placed
in a specially designed vacuum chamber(3) where the filters are irradiated
with 5 MeV protons from FSU1 s super-FN Tandem Van de Graaff. (A small verti-
cal Van de Graaff dedicated to PIXE,-analysis is planned for operation in 1979
and will allow the production of 10 data points a year). The x-rays gener-
ated are detected using an energy dispersive Si(Li) detector purchased from
Nuclear Semiconductor, which has a resolution of about 170 eV FWHM for the
iron K x-ray (6.4 keV). The detector signals are processed using Nuclear
Semiconductor electronics which include an amplifier, a pulse pile up rejec-
tor, a low level discriminator, and live time counter. Processed signals are
fed into a Northern Scientific ADC which is Camac interfaced to an EMR 6130
on-line computer. At the end of each run the data is written on magnetic tape
for off-line analysis. The PIXE technique is well understood(4) and the Ele-
mental Analysis Group has taken part in numerous intercomparison studies.(5,6)
A sophisticated FORTRAN code REX(7) is used to fit the x-ray spectra.
This code carefully models the spectra obtained. It contains information re-
levant to elemental x-ray line shapes, the Bremsstrahlung background, the si-
licon absorption edge, x-ray production cross sections, self absorption and
attenuation, and other parameters. Absolute calibration is made from stand-
ards produced by Micro-Matter, Inc. and checked by PESA.(2)
During this experiment flow rates were maintained using a Mass Flow Con-
troller TYLAN Model FC 260. The precision of this device was confirmed before
and after the experiment using a spirometer and a factory calibrated variable
area flowmeter (Matheson Model 601).
As a side experiment two other linear streakers were run. One was uncon-
trolled at 0.8 t/m and one had a prefilter of 8 jjm pore size. Modified Bat-
telle cascade impactors were used as well with 4 hour sampling periods. These
impactors have size cuts > 4, 4-2, 1-0.5, 0.5-0.25 and < 0.25 m, respectively.
Mylar impaction surfaces and 0.4 urn Nuclepore afterfilters were employed.
Petroleum jelly was used to coat the impaction surface except for the last
impaction stage which was coated with parafin. Flow rates of the impactors
were controlled by a critical flow orifice and were measured with a calibrated
variable area flowmeter. Additional results from these samplers will be re-
ported in a separate publication at a later date.
REFERENCES
1. Novel Aerosol Sampling Apparatus for Elemental Analysis, B. Jensen and
J. W. Nelson, Proceedings of the 2nd International Conference on Nuclear
Mehtods in Environmental Research - U. of Mo., July, 1974, US AEC Conf. -
740701, 385, Oak Ridge, Tenn.
2. Light Element Analysis by Proton Scattering, J. W. Nelson and W. J.
Courtney, Nuclear Instruments and Methods 142, 127 (1977).
3. Proton Induced Aerosol Analysis: Methods and Samples, J. W. Nelson, X-ray
Florescent Analysis Environmental Samples, Editor Thomas G. Dzubay, Ann
Arbor Science Publishers, Inc. (1977).
82
-------
4. Analytical Application of Particle Induced X-Ray Emission, Sven A. E.
Johansson and Thomas B. Johansson, Nuclear Instruments and Method's 137,
473 (1976).
5. D. C. Camp, A. L. Van Lehn, J. R. Rhodes, A. H. Pradzynski, J. X-Ray
Sepctr. 4, 123 (1975).
6. Continuous Observation of Particulates During the General Motors Sulfate
Dispersion Experiment, W. J. Courtney, S. Rheingroves, J. Pilotte,
H. C. Kaufmann, T. A. Cahill, and J. W. Nelson, Jour. APCA, March 1977.
7. Rex - A Computer Programme for PIXE Analysis, Henry C. Kaufmann,
K. Roland Akselsson and William J. Courtney, Nuclear Instruments and
Methods 142, 251 (1977).
83
-------
AUTOMATED DICHOTOMOUS SAMPLERS
B. W. Loo
Lawrence Berkeley Laboratory.
Berkeley, CA 94720
QUALITY ASSURANCE MEASURES
Samp!i ng
1. Flow calibrations performed in the field.
2. Automated flow controller in the sampler maintains the pre-
set flow rate.
3. Fault sensors continuously check for filter loading, proper
filter transport and vacuum seals.
4. Flow calibrations checked upon completion of study.
Analysis
1. Checked calibration and XRF spectrometer gain setting.
2. Checked helium level during XRF analysis.
3. Calibrate beta-gauge every hour.
4. Check and correct for long term beta-gauge response.
5. All samples measured twice for both total mass and elemental
composition which checks for precision and any obvious error.
DETAILED ANALYTICAL METHODS AND PROCEDURES
Mass Measurement
1 47
A beta gauge consisting of a Pm source and a cooled Si detector was
used. The area! mass of all filters was measured twice prior and subsequent to
to field sampling. Ihe measured beta intensity I was related to the sample
thickness X in jjg/cm by I = I.e'"1* where IQ and jj were constants derived from
a leastsquare fit to five calibration standards. These standards were gravi-
metrically weighted polycarbonate films of thicknesses 1034.6, 1511.9, 1970.6,
2552.2 and 2931.0 prg/cm . The counting times used for calibration and rou-
tine measurement were 100 and 30 sec. respectively. About 4 x 10 counts were
accumulated within 30 sec. for a sample of 1 mg/cm . For a mean aerosol con-
centration of 107/jg/cm the mean reproducibility was 4.7 jug/cm . A correc-
tion factor for long term systematic changes of 12.5 jug/cm was determined
from the remeasurement of 25 unexposed filters.
84
-------
XRF Measurements
An XRF spectrometer including a pulsed x-ray source, cooled anti-coin-
cidence guard-ring Si(Li) detector and fluorescent targets of Ti, Mo and Sm
were used for elemental analysis. The x-ray calibration factors were deter-
mined using a series of thin-film standards which closely replicate the mem-
brane filter samples. These standards are either uniform evaporation deposits
of single element whose mass has been determined gravimetrically or multiple
element standards in which the elemental concentration ratios are accurately
known. For example the sulfur calibration was obtained by measuring specimens
of 0.3 jjm Cu S04*5H20 aerosols deposited on the surface of 0.1 jum pore size
nucleopore filters. In this case evaporated deposits were used as primary
standards. This calibration was cross-checked with that determined from a
series of relative standards, namely KpS04, K2Cr20 and Cu-Cr mixed aerosol
standards. The agreement was within 1%.
The x-ray attenuations due to the size of the particles were corrected
using the method described by Dzubay and Nelson (Adv. in X-ray Anal. 18, 619,
1975). The values employed are listed as follows:
Element Fine Particles Coarse Particles
Si
S
Ca
Ti
Fe
Cu
Zn
0.93 + 0.07
0.97 + 0.03
0.99 + 0.01
0.48 + 0.15
0.64 -(- 0.22
0.81 + 0.13
0.87 + 0.10
0.94 + 0.05
0.94 + 0.06
0.95 + 0.05
Due to the flow division within the virtual impactor, the reported values on
size segregated samples have an implicit 5% interference of fine particle on
the coarse particle fraction. This effect, of course, disperses when the two
size fractions are summed. The only case where it may introduce any signifi-
cant error is where the particles are found predominantly in fine particles
and where the coarse particle size attenuation correction is large. This was
the case with S and the attenuation corrections were applied after making the
5% fine particle interference correction.
85
-------
CYCLOME SEQUENTIAL FILTER SAMPLERS
P. K. Mueller, D. A. Hansen and S. L. Heisler
Environmental Research & Technology, Inc.
2625 Townsgate Road
Westlake Village, California 91361
The samplers used in the aerosol sampler intercomparison co^KRed in
Charleston, West Virginia between May 11 and May 19, 1977, were ^RrTModel 5200
Sequential Filter Samplers. The sampling head consists of a 5 cm I.D. by 50
cm long tube topped by a cylindrical shroud. The purpose of the shroud is to
render the entry of large particles into the sampler relatively insensitive to
changes in wind speed. The conical covering prevents entry of rain. Wire
screens are located over the top and the bottom of the shroud to prevent entry
of insects into the duct.
One of the samplers, F, was equipped with a cyclone preseparator located
within the shroud. The Unico 240 cyclone used was calibrated for particle re-
moval by Eisenbud and Kneip (1975). Based on their results, the 50 percent
collection efficiency size for unit density spheres varied from 2.0 to 2.7 pm
diameter as the flow rate varied over the range from 100 to 150 4pm. The sam-
pler equipped with the cyclone collected only particles smaller than these
sizes. The second sampler did not have a preseparator and collected particles
smaller than approximately 15 urn diameter.
Sampled air passes through the inlet duct to the cylindrical chamber
containing position for 13 filter holders. The circular arrangement of the
filters subjects each to similar flow patterns. Air is never sampled through
the thirteenth position; a filter is placed in this position to assess the ef-
fects of particle deposition when sampling is not occurring.
The filter holders are connected to a common pump through a system of
solenoid valves. The valves are actuated by timers to allow sampling of the
twelve filters in succession. The sampling times are programable for periods
from 5 minutes to 24 hours by means of a 24-hour time switch. A sampling se-
quence can be programmed to start at any time within seven da'ys after set-up.
The deviation of the actual starting time from the programmed time is 30 min-
utes or less.
Flow rates are monitored continuously by means of a vacuum recorder con-
nected upstream of the pump. The recorder is calibrated for flow rate.
The samplers accommodate open-face 47 mm diameter filter holders. The
filter holders are equipped with quick-disconnect type fittings to allow easy
installation and removal. Filters were shipped to and from the field in fil-
86
-------
XRF Measurements
An XRF spectrometer including a pulsed x-ray source, cooled anti-coin-
cidence guard-ring Si(Li) detector and fluorescent targets of Ti, Mo and Sm
were used for elemental analysis. The x-ray calibration factors were deter-
mined using a series of thin-film standards which closely replicate the mem-
brane filter samples. These standards are either uniform evaporation deposits
of single element whose mass has been determined gravimetrically or multiple
element standards in which the elemental concentration ratios are accurately
known. For example the sulfur calibration was obtained by measuring specimens
of 0.3 jum Cu SO^'S^O aerosols deposited on the surface of 0.1 jum pore size
ntfccleopore filters. In this case evaporated deposits were used as primary
standards. This calibration was cross-checked witb that determined from a
swies of relative standards, namely KoSO^, Kfirfi and Cu-Cr mixed aerosol
standards. The agreement was within it.
The x-ray attenuations due to the size of the particles were corrected
using the method described by Dzubay and Nelson (Adv. in X-ray Anal. 18, 619,
1975). The values employed are listed as follows:
Element Fine Particles Coarse Particles
Si
S
Ca
Ti
Fe
Cu
Zn
0.93 + 0.07
0.97 + 0.03
0.99 + 0.01
0.48 + 0.15
0.64 + 0.22
0.81 + 0.13
0.87 + 0.10
0.94 + 0.05
0.94 + 0.06
0.95 + 0.05
Due to the flow division within the virtual impactor, the reported values on
size segregated samples have an implicit 5% interference of fine particle on
the coarse particle fraction. This effect, of course, disperses when the two
size fractions are summed. The only case where it may introduce any signifi-
cant error is where the particles are found predominantly in fine particles
and where the coarse particle size attenuation correction is large. This was
the case with S and the attenuation corrections were applied after making the
5% fine particle interference correction.
85
-------
CYCLOME SEQUENTIAL FILTER SAMPLERS
P. K. Mueller, D. A. Hansen and S. L. Heisler
Environmental Research & Technology, Inc.
2625 Townsgate Road
Westlake Village, California 91361
The samplers used in the aerosol sampler intercomparison conducted in
Charleston, West Virginia between May 11 and May 19, 1977, were ERT Model 5200
Sequential Filter Samplers. The sampling head consists of a 5 cm I.D. by 50
cm long tube topped by a cylindrical shroud. The purpose of the shroud is to
render the entry of large particles into the sampler relatively insensitive to
changes in wind speed. The conical covering prevents entry of rain. Wire
screens are located over the top and the bottom of the shroud to prevent entry
of insects into the duct.
One of the samplers, F, was equipped with a cyclone preseparator located
within the shroud. The Unico 240 cyclone used was calibrated for particle re-
moval by Eisenbud and Kneip (1975). Based on their results, the 50 percent
collection efficiency size for unit density spheres varied from 2.0 to 2.7 \im
diameter as the flow rate varied over the range from 100 to 150 4pm. The sam-
pler equipped with the cyclone collected only particles smaller than these
sizes. The second sampler did not have a preseparator and collected particles
smaller than approximately 15 urn diameter.
Sampled air passes through the inlet duct to the cylindrical chamber
containing position for 13 filter holders. The circular arrangement of the
filters subjects each to similar flow patterns. Air is never sampled through
the thirteenth position; a filter is placed in this position to assess the ef-
fects of particle deposition when sampling is not occurring.
The filter holders are connected to a common pump through a system of
solenoid valves. The valves are actuated by timers to allow sampling of the
twelve filters in succession. The sampling times are programable for periods
from 5 minutes to 24 hours by means of a 24-hour time switch. A sampling se-
quence can be programmed to start at any time within seven days after set-up.
The deviation of the actual starting time from the programmed time is 30 min-
utes or less.
Flow rates are monitored continuously by means of a vacuum recorder con-
nected upstream of the pump. The recorder is calibrated for flow rate.
The samplers accommodate open-face 47 mm diameter filter holders. The
filter holders are equipped with quick-disconnect type fittings to allow easy
installation and removal. Filters were shipped to and from the field in fil-
86
-------
ter holders to eliminate the need for filter handling at the sites. A protec-
tive plastic cap is placed over the filter holder during shipment and handling,
The vacuum recorders in the samplers were calibrated three times during
the program: on 10, 15, and 19 May 1977. Calibrations were performed with
dry test meter attached at the sampler inlet. The cyclone preseparator was
removed from the RSP sampler and the dry gas meter was connected in its place.
Various types and combinations of filters were installed in the sampler to
produce various flow rates for the calibration. The calibration was performed
by cycling the sampler through these filters and recording the readings of the
dry gas meter and the vacuum recorder. In the case of the TSP sampler which
did not have a cyclone, the cylindrical chamber from the RSP sampler was in-
stalled for the calibration.
The results of the first and second and the second and third calibrations
for each sampler were taken together and linear least-squares fits to each set
of two calibrations were made. The calibration curve from the first set was
used for reducing the data from the 11-14 May sampling; that from the second
set was used for the 15-19 May sampling data.
CALCULATIONS OF CONCENTRATIONS
To calculate the volume of air sampled during each 3-hour sampling inter-
val, the initial and final flow rates were averaged and multiplied by the time
clasped during the interval.
The quantity of anionic material on each filter was determined as the
product of the concentration of the anion in the extracted solution times any
dilution factor times the reciprocal of the fraction of the filter which was
extracted. These products, in nanograms, for four 3-hour intervals were added
and the sum was divided by the sum of the volumes of air sampled during the
four intervals to yield the average concentration of the anion in air, in ng
m , for each 12-hour reporting period.
FILTER MEDIUM
Particulate samples were collected on 47 mm diameter Teflon-impregnated
glass fiber disks with the designation TX40HI20 manufactured by Pallflex Pro-
ducts Corp. to ERT specifications. This choice was based on the following
considerations:
t The filter medium must retain particles efficiently while transmitting
gases.
• Flow rates achievable using the filters in the Sequential Filter Sam-
plers must be sufficient to allow collection of adequate material for
analysis by the methods chosen.
• The filters must hav/e sufficient mechanical strength for easy handling.
The first consideration, transmission of gases, eliminated ordinary glass fi-
ber filters, because filters of this type have been shown to interact strongly
with sulfur dioxide and other acid gases such as nitrogen oxides leading to
erroneously high measurements of sulfates and nitrates. Membrane filters were
87
-------
eliminated by the second consideration because of their high flow resistance.
The remaining types of filters to be considered were quartz fiber filters and
the Teflon-impregnated glass fiber filters. Evaluation of the quartz fiber
filters indicated they were too fragile to handle without significant loss of
filter material.
GRAVIMETRIC ANALYSIS
Weighings were performed with a Cahn Model 4700 Automatic Electro^balance.
Static charges were removed from the balances and from filters with Po ion-
ization units located inside the weighing chambers in the balances.
Filters were equilibrated for a minimum of 24 hours in the balance room
prior to weighing. The room was maintained between 23 and 25°C and at a rela-
tive humidity below 50 percent. Unexposed filters were equilibrated in the
containers with the lid open in which they were received from the manufacturer.
Exposed filters remained in the shipping containers with the lid open and the
protective caps in place during equilibration.
Filters were inspected visually for defects prior to weighing. Unexposed
filters showing defects were discarded.
The balance was operated at a full-scale range of 2000 pg with a resolu-
tion of 10 ug. Calibration was performed with Class M weights.
ANION ANALYSIS
Nitrate and sulfate were determined from a single extract by the techni-
que developed by Small and co-workers (1975). This technique is called ion
chromatography (1C).
A quarter of the filter was placed in a numbered borosilicate culture
tube which had been previously cleaned and conditioned with eluent buffer,
and 10 ml of buffer were added by means of a Repipet liquid dispenser. The
tube was then capped and shaken ultrasonically or minutes in a Bransonic 32
ultrasonic cleaner. Because standing waves may exist at some points in the
cleaner, the tubes were moved continuously during the extraction process.
Stock solutions of standard were prepared by dissolving analytical grade
reagents in distilled deionized water and storing them in the refrigerator.
Standard mixtures of NOo and SO^were prepared in eluent buffer daily. All
solutions were allowed to equilibrate at the instrument room condiitons before
analysis.
A Model 10 ion chromatograph (Dionex Corp., Palo Alto, CA) was used for
the analysis of the anions. For maximum resolution of the ions of interest,
the instrument was operated at a pump speed of 2 ml/min. The eluent buffer
was 0.001M Na2C03 and 0.003M NaHC03 prepared in distilled deionized water
having a conductivity of less than 0.1 jumho/cm. All samples were run at the
10 umho full-scale setting.
A disposable syringe was used to load the solution into the sample loop.
88
-------
Each sample was individually filtered in-line with a 0.2 ^m pore size membrane
filter during the sample-loading step in order to prevent contamination of the
column with insoluble matter. Injections of the samples were spaced by fif-
teen minutes to allow complete elution of all anions.
For each ionic species, the peak height was measured from the defined
baseline of initial to final inflection.
Calibration curves were prepared from a plot of peak height in inches
versus concentration in ug/ml. The linear regression for each curve was used
to reduce the data.
89
-------
HIGH VOLUME AND OICHOTOMOUS SAMPLERS
Charles E. Rodes
Environmental Monitoring and Support Laboratory
Enivronmental Protection Agency
Research Triangle Park, NC 27711
INTRODUCTION
The Environmental Monitoring and Support Laboratory (EMSL) of EPA will
participate in the Aerosol Sampler Comparison Study by operating 2 high-vol-
ume (hi-vol) samplers and 2 EMSL dichotomous samplers. EMSL will be respon-
sible for all aspects of our portion of the study including: (1) equipment
and personnel transportation, (2) set-up and take-down, (3) oepration during
the 8 day period and (4) analysis for selected parameters.
SAMPLING DEVICES AND FILTER SUBSTRATES
The hi-vols used will conform to the standard Federal Register (1) re-
quirements with the addition of constant mass flow controllers and Dixon flow
recorders. The hi-vols will be operated at 1.42 snr/min (50 SCFM) using acid
washed 200 x 250 cm (8 x 10 in.) glass fiber filters.
The EMSL dichotomous samplers have the standard "Coarse" and "Fine"
channels with a 3.5 urn cutpoint and an inlet cut-off of ^20 urn. These sam-
plers also include a separate "Total" channel which collects aerosols < 20 urn.
Flow control is maintained by critical flow valves with rotameters on all 3
channels. The sampler utilizes 37 mm Fluoropore Teflon filters with a poly-
ethylene backing. These filters are backed with a 37 mm cellulose pad and
placed in circular holders which fit holders in the respective sampler chan-
nels.
QUALITY CONTROL
The samplers will be checked for flow calibration only. All samplers
will be calibrated initially in RTP using a Rootsmeter® as the traceable
standard. After the samplers are set up in the field, transfer flow stand-
ards will be used to audit check the flowrate initially and in the middle of
the study. A standard audit orifice also calibrated using the Rootsmeter® is
used to check the hi-vols. A 300 mm rotameter calibrated in the lab against
a dray-test meter is used to audit check the "Total" and "Fine" dichotomous
channels. The "Coarse" channel is checked with a calibrated mass flowmeter.
At the conclusion of the study all samplers will be recalibrated with the
audit flow devices.
90
-------
During sample collection a change in flowrate from the beginning to
the end of a sampling period (12 hrs) of + 15% is allowable. Flow changes
greater than this would invalidate the respective sample.
In order to obtain reproducibility information 2 samplers of each type
will be operated simultaneously for 8 sample periods (4 days). After these
comparison tests the samplers will be operated alternately to obtain only one
sample of each type per time period.
ANALYSES
The primary measurements will be gravimetric mass using a 5-place ba-
lance, with subsequent analyses for sulfate (S0^~) by the MTB method and lead
(Pb) by atomic absorption. Mass measurements will be made on all fiters.
Two measurements - SO*" and Pb - will be performed on all samples collected
except the duplicate aichotomous samples. The wet chemistry and mass analy-
ses will be performed by the Analytical Chemistry Branch of EMSL.
REFERENCES
1. Federal Register, Vol. 36, No. 84, April 30, 1971, pp. 8191-8194.
91
-------
HIGH VOLUME AND DIFFUSION SAMPLER
Roger L. Tanner and Leonard Newman
Atmospheric Sciences Division
Brookhaven National Laboratory
Upton, New York 11973
The samplers used in this study are a conventional Staplex High Volume
sampler sampling at about 0.9 m /min and a diffusion sampler sampling
through three 47 mm filters at a total flow rate of about 2.5 m /hr. The Hi-
Vol sampler samples all particles less than about 20 pm in diameter. All
three diffusion-processed samples were taken through an entrance tube that
excluded most particles larger than about 5 urn in diameter: Sample DBO sam-
ples all remaining size ranges of particles; particles smaller than about
0.04 jjm are removed by diffusion from sample DB1, particles smaller than
about 0.1 urn are removed from sample DB2.
Samples from both the Staplex HiVol and the diffusion sampler were col-
lected on phosphoric acid-treated quartz filters (1) (Pall Corp., Putnam,
Conn., Type GAO). This filter material is prepared as follows: the filter
as received is ignited overnight at 750°C to remove binder material; filter
is then washed then treated with hot (80°C) phosphoric acid; following a
rinse to remove excess H3P04, filter is reignited to 750°C. A post-ignition
rinse in Ph 5 HC1 removes excess ?2®5 anc' following drying at 100°C, the fil-
ter is ready for use. The HiVol sampler was also used to collect an S02 sam-
ple during each period on two 1^03, glycero-impregnated cellulose filters in
series with the quartz prefilter.
The HiVol sampler was mounted in a conventional HiVol enclosure pro-
tected from precipitation and could be operated under any precipitation con-
ditions. The diffusion-sampling pump was placed in a protective box but in
the absence of an enclosed shed, no protection was used with the diffusion
battery and filter holders; although no diffusion sampling during heavy rain
conditions could be conducted, these conditions were not encountered during
the study.
The HiVol sampler has been calibrated with the filter pack in place
using a magnetohelic guage. Decreses in flow during 12 hr sampling were
checked out after the experiment and found to be less than 1 cfm, insignifi-
cantly low to require measurement for each sample. The diffusion sampler has
been calibrated using unrestricted Matheson rotameter tubes in series up-
stream from the three 47 mm filter holders, (i.e., in line at ambient pres-
sure). This calibration was assumed to hold constant for 47 mm quartz
filters from the same batch of treated quartz. The calibration was checked
after the experiment for constancy; changes of only a few % were observed
92
-------
and the average of flows before and after the Intercomparison were used for
sample volume calculations.
Each 4-in diameter, HiVol quartz particulate filter was quartered with
one filter quarter sequentially extracted by benzaldehyde and isopropanol and
the extract contents analyzed for HpSO^ and bisulfate, respectively, by the
flash volatilization-FPD techique (3). A second quarter was extracted into
pH 4 aqueous solution and soluble strong acid determined by Gran titration
(1), ammonium and nitrate ions by Autoanalyzer colorimetric techniques (1),
and soluble sulfate determined by ion chromatography (4) and also by Methyl -
thymol Blue colorimetry (5). The reamining quarters were saved for replicate
determinations as needed or for cross checks of sulfur content by the reduc-
tion-sliver-110 technique (2). The SOp was determined in the two I^CC^,
glycerol-impregnated cellulose backup filters by extraction into aqueous
peroxide with determination as sulfate by ion chromatographic or turbidime-
trie Autoanalyzer techniques.
The diffusion-processed samples, collected on 47 mm treated quartz +
filters, were bisected. One half was analyzed for soluble strong H , NH/^,
$04 and NO^ by the methods described above; the second half was sequen-
tially extracted by benzaldehyde for determination of ^SO^ and also by the
techniques described above.
REFERENCES
(1) R. L. Tanner, R. Cederwall, R. Garber, D. Leahy, W. Marlow, R. Meyers,
M. Phillips and L. Newman, "Separation and Analysis of Aerosol Sulfate
Species at Ambient Concentrations", Atoms. Environ., Y\_, 955 (1977).
(2) J. Forrest and L. Newman, "Application of Ag Microgram Sulfate Ana-
lysis for the Short Time Resolution of Ambient Levels of Sulfur Aerosol",
Anal. Chem., 49, 1579 (1977).
(3) R. L. Tanner, R. W. Garber and L. Newman, "Speciation of Sulfate in
Ambient Aerosols by Solvent Extraction with Flame Photometric Detection",
173rd National Meeting, American Chemical Society, New Orleans, LA,
March 20-25, 1977.
(4) W. E. Rich, "Analysis of Nitrate and Sulfate Collected on Air Filters",
Application Note No. 2, Dionex Corp., Palo Alto, CA, January 23, 1976.
(5) J. M. Adamski and S. P. Villard, "Application of the Methyl thymol Blue
Sulfate Method to Water and Wastewater Analysis", Anal. Chem., 47, 1191
(1975). ~
93
-------
APPENDIX B
OUTLIER AND REGRESSION ANALYSIS EQUATIONS
OUTLIER TEST
The Dixon criterion is a statistical test applied to any set of n obser-
vation, where n > 3. Its objective is to identify an observation, which to
within a chosen probability, does not belong to the other (n -1) observations.
These do not include observations where there is a known cause for its devia-
tion. Otherwise when the deviation is larger than one can reasonably expect,
this should serve as a stimulus to discover what happened. If it appears that
nothing happened then the results in this intercomparison study were left as
is; however, the outlying values were flagged for all to see.
Let n be defined as the number of real non-zero sampler concentrations
reported for a given period of the intercomparison study. Four coefficients
are defined by each periods result set; however, only one coefficient will
apply. Its use and definition depend on the number of results for that pe-
riod, and whether the suspected observation lies at the low or high end. If
the observations are ordered from 1 to n, where the first value, X^, is the
lowest concentration reported and X_ is the greatest, then the four coeffi-
cients are defined as follows depending on n.
rl: 3 1 n < 7
r2: 8 < n < 10
r3: 11 < n < 13
r4: 14 < n < 25
If result Xn is suspect, then
r, = IV - Y
rl _ un An
r4 = (Xn - Xn
If the result X1 is suspect, then
_
- X2)
- X2)
- Xo)
'\: 8f:
r3 =
r4 =
- Xi)
(B-l)
(B-2)
(B-3)
(B-4)
(B-5)
(B-6)
(B-7)
(B-8)
(B-9)
(B-10)
(B-ll)
(B-12)
The rj coefficients are calculated and compared to tabulated coefficients
r(l-a/2), where a is defined as the probability assumed in rejecting a re-
94
-------
suit that really belongs in the group. For the intercomparison study a was
defined as 4%, thus any value greater than 2% high or less than 2% low would
be defined as an outlier.
The calculated r^ coefficients are compared to standard tables, a portion
of which is given in Table B-l. Although e* was chosen as 4%, if « is chosen
smaller the coefficients will become larger and the likelihood of rejection
decreases.
TABLE B-l. CRITERIA FOR REJECTION OF OUTLYING OBSERVATIONS
Upper Percentiles
Coeffi-
cient
ri
r2
# of
Observ.
3
4
5
6
7
8
9
10
.95
.941
.765
.642
.560
.507
.554
.512
.477
a =4
.98
.976
.846
.729
.644
.586
.631
.587
.551
a =2 Coeffi-
.99 cient
.988
.889 r3
.780
.698
.637
.683 r4
.635
.597
Upper Percentiles
# of
Observ.
11
12
13
14
15
16
17
0.95
.576
.546
.521
.546
.525
.507
.490
a =4
0.98
.638
.605
.578
.602
.579
.559
.542
a =2
0.99
.679
.642
.615
.641
.616
.595
.517
More detailed information on the Dixon criterion can be found in Experimental
Statistics, a National Bureau of Standards Handbook 91, Chapter 17, by
M. G. Natrella, issued August 1, 1963, reprinted October 1966, with correc-
tions.
REGRESSION ANALYSIS
Assume that the pollutant concentration results obtained for sampler Y
are linearly correlated with those obtained for sampler X. Let j be the per-
iod number (j = 1, 2, 3,..., 16). Then for any sampler Y we can write
(B-13)
y = a + bx
The intercept, a, will be given by
a =
a n
SXX
[B-14)
and
b =
SXY
SXX
_ (IX)
(B-15)
95
-------
and n is defined as the number of paired XjY, values (< 16). Note that when
regression analysis is performed between any sampler and the composite Z-
means there will always be 16 Z- values, but not necessarily 16 paired Y-Z.
values. J J
Similarly, when regression analysis is performed between any two samplers
, the number of paired values will define n. The errors ea and eb can al-
; calculated using
ea =
i (TYT -
(SXY)
SXX
r
Ln
(X)2-,
SXX J
(B-16)
where
_l __ L
n-2 SXX
Y - —
Y n
SXY
SXX
1
2
(B-17)
the
The root mean square residual, r, of the fit is one measure of how well
two samplers define a linear relationship. The results from two samplers may
tightly define their mutual slope, b, in which case r is small. However, if
two samplers report widely fluctuating values, then r will be large. If all
samplers are compared to the composite sampler, then an intercomparison of
their respective r1 s is a measure of their relative precisions (assuming that
b is approximately 1.0). The root mean square residual concentration for sam-
pler Y relative to sampler X is defined by
r =
i m
*£
(YJ
1.
^ o o O O
- Y) ] in ng/m (or pg/m )
(B-18)
where there are m concentration results for sampler Yj and Y is given by
(B-13). Finally, it is useful to examine how well the two samplers X and
Y are correlated, i.e., how linearly related there concentrations are.
The correlation coefficient is given by
(bXY ' bYX)
1
SYX.?
'
SYY
(B-19)
SXY was given in (B-15) and SYX is obtained by interchanging X and Y in that
expression.
The intercept and its error, the slope and its error, the root mean
square deviation in ug or ng/m , and the correlation coefficient are listed in
Appendix C for each sampler that reported results for any pollutant fraction
or total concentration.
96
-------
More detailed information on linear regression analysis of two variables
can be found in Experimental Statistics, a National Bureau of Standards Hand-
book 91, Chapter 5, by M. G. Natrella, issued August 1, 1963, reprinted Octo-
ber 1966, with corrections.
97
-------
APPENDIX C
Table of Contents
Introduction, Definitions, and Comments 100
Pollutant Total/Small/Large
Mass 101
Nitrate 104
Sulfate 107
Sulfur 110
Sulfur + Sulfate 113
Silicon 116
Calcium 119
Titanium 122
Iron 125
Copper 128
Zinc 131
Selenium 133
Bromine 136
Lead 139
99
-------
INTRODUCTION DEFINITIONS COMMENTS
POLLUTANT RESULTS ARE GROUPED TOTAL, SMALL, 8. LARGE WITH THE FOLLOWING ORDER:
MASS, N03, SOH, S, S+SOH, 8, TRACE ELEMENTS-SI, CA. TI, FE, CU, ZN, SE, BR, 8. PB.
TOTAL:REPORTED TOTAL CONCENTRATIONS OR SUM OF ALL FINE 8. COARSE FRACTION RESULTS
SMALL: USUALLY THE FINE FRACT10N(S); CUT POINTS WILL VARY WITH SAMPLER TYPE:
SMALL MEANS ABOUT 3.0 + OR - 0.5 MICROMETERS AND SMALLER
LARGE: THE COARSE FRACTION; INLET RESTRICTIONS AND FINE CUT POINTS VARY ALSO.
LARGE MEANS ABOUT 3.0 + OR -0.5 MICROMETERS AND GREATER, UP TO THE INLET CUTOFF
MASS RESULTS 8. ERRORSINOT SHOWN) ARE PUNCHED IN TENTHS OF MICROGRAMS/CUBIC METER
BUT ROUNDED OFF IN THE TABLES. ALL OTHER RESULTS ARE LISTED IN NANOGRAMS/CUB1C M
SAMPLER DEFINITIONS
DICOT= AUTOMATED DICHOTOMOUS SAMPLERS (3-C. L, 8. S) FROM LBL-LOO
DICOT= AN ERC DICHOTOMOUS SAMPLER (1-0) FROM EPA-RODES
HIVOL= HIGH VOLUME SAMPLERS (3: A FROM BNL-TANNER 8> P/T FROM EPA-RODES/BURTON)
LS = LINEAR STREAKER SAMPLERS (2 - J 8, K) FROM FLA .ST .UNI V. - HUDSON
MDS = MANUAL DICHOTOMOUS SAMPLERS (2 - D & R) FROM EPA-DZUBAY
RSP = CHAMP SAMPLERS (2-B8.U) FROM ERA-BURTON; CYCLONE SAMPLER!I-r) FROM MUELLER
SMDAY^ SIERRA MULTIDAY SAMPLER (1 - G) 3-STAGES UNI V.CA.DAVIS-CAHILL
SFU = STACKED FILTER UNIT (1 - I) 2-STAGES FROM UN IV.CA.DAVIS-CAHILL
TTLF = TOTAL FILTER UNIT (1 - H) 1-STAGE FROM UNI V.CA.DAVIS-CAHILL
TSP = CYCLONE SAMPLER (1 - E) FROM ERT-MUELLER
2MASS= TWO MASS SAMPLER (1 - Q) FROM WASH.UN I V. ,ST.LOUIS-DELUMYEA/MACI AS
FOR EACH POLLUTANT, THE FIRST RESULT TABLE LISTS THE REPORTED CONCENTRATIONS AS
A FUNCTION OF PERIOD, EACH 12 HOURS LONG, AND LISTED FROM COLUMNS 1 THROUGH 16.
EACH SAMPLER IS ASSIGNED AN ALPHABETICAL INDEX ACCORDING TO ITS SPATIAL LOCATION
AT THE FIELD STUDY. SEE FIGURE 1 AND TABLE H IN THE 'EXT FOR MORE INFORMATION.
A DASH - ENTERED IN THE TABLE INDICATES NO VALUE REPORTED FOR THAT SAMPLER 8. PD.
OUTLIER VALUES ARE DENOTED BY A -(NEC) SIGN BUT ARE INCLUDED IN MEAN CALCULATION
THE Z-MEAN INCLUDES ALL SAMPLER RESULTS & OUTLIERS EXCEPT THOSE IN PARENTHESIS.
THE X-MEAN INCLUDES ONLY THOSE SAMPLER RESULTS NOTED IN PARENTHESIS-NO DUPLICATE
THE SECOND TABLE LISTS RATIOS OF REPORTED RESULTS TO THE APPROPRIATE Z-MEANS.
THE LAST 3 COLUMNS LIST THE 16-PERIOD AVERAGE RATIOS, THEIR STANDARD DEVIATIONS,
AND THE PERCENT ERROR ASSOCIATED WITH THAT STD.OEV.(COEFFICIENT OF VARIATION)
AN ASTERISK BY ENTRIES IN THIS TABLE INDICATE THAT A PARTICIPANTS REPORTED ERROR
FOR THAT PERIOD ALLOWS HIS RESULT TO OVERLAP 1.0 OR THE Z-MEAN CONCENTRATION.
NOTE THAT THE STANDARD DEVIATIONS VS PD. IN THE MEAN RATIOS INDICATE THE DEGREE
OF SCATTER IN THE REPORTED CONCENTRATIONS FOR THAT PARTICULAR PERIOD 8, POLLUTANT
THE LAST TABLE LISTS THE RESULTS OF A LINEAR REGRESSION ANALYSIS OF EACH SAMPLER
WITH THE Z-MEAN. LEFT TO RIGHT ARE THE INTERCEPT,ITS ERROR; THE SLOPE,ITS ERROR;
THE ROOT MEAN SQUARE DEVIATION IN CONC. UNITS; AND THE CORRELATION COEFFICIENT.
-100-
-------
Q
I/I
Q
cn
— inp-cooD3-
incoin3-cOf*icocnccp-o>
OCOOO — OOOOM
r-
cn
CO
'
X
CO
C/l
8
u
D.
OC
c_>
o
iO
•IVHOOM:
z:
in
*
'
H-
LL.
O
O
h-
cr
g
o
or
t—
z
o
co
in
3-
ro
ru
-
0
cn
CO
p-
co
in
3-
ro
ru
—
RESULTS PERIOD
0
o
3"
1
O
en
in
r-
cn
to
CO
r
i
o
&s
1
cn
p-
S
cn
o
p-
i
l/i
o
r
0
<
Q.
U)
O
CC 0
CO •=>
o ru
in ro
ru ro
cn co
- ru
co co
01 —
ru p-
03 Ol
O CO
CD cn
i ro
CO
ru p*
3- 3-
ro i
3- 0
cn —
CD p-
00 —
in co
co o
— CD
p- cn
O 1
p-
S '
1 O
in u
Q _
r a
D O
M O
C _l
< _l
a, co
Ul _J
rr o
_i in
ru CD
ro ru
ru cn
3- ro
ru p-
ro ro
CO I
ru
CO i
cn
CD i
o
3- P-
P- P-
0 CO
in 3-
i i
in P-
— ru
ru CD
ru —
ro 3-
- ru
p- i
CD
in en
U) t— '
in CD
cn cn
CO 01
co cn
r— t—
O O
CJ CJ
Q O
0 O
0 0
m co
_l C/l
CL
p-
ru
ro
CO
p-
CO
in
ru
co
ru
cn
CD
in
ro
3"
ro
cn
ru
cn
in
co
ro
CD
ro
~
f
>
I
cn
Ul
X
a.
CL
CD i-
P- CO
3- ro
cn p-
m r-
01 CO
3- in
— o
ro m
ro ru
— ru
0 3-
~ ru
ru in
CD cn
r> o
in in
3- o
r- ru
o 3-
ru ru
ru —
ru CD
cn in
CO —
r- p-
p *
oo£
PI .
ru cn
cn —
B EPA BURTON CHAMP
T EPA BURTON HIVOL
^ UJ
cn 3-
in in
10 ru
cn cc
ioS
00 0
— CO
3- CO
o CO
— ru
co m
CO 3-
3- ro
ru p-
ru —
- O!
— cn
ru cn
10 in
ru co
01 ru
O> CD
3- cn
cn 3-
CD P-
3- ro
P- CD
a.
< a.
i in
O 1-
z rr
O Ul
oo r
0- CC
Ul Ul
O ul
0
1
<
p-
in
-
i
i
ru
cn
i
i
i
I
I
S
r-
ro
in
~~
cn
G STL DEL/ED 2MASS
j-
o
in
ru
•
CO
3-
CD
ro
*
ru
ru
o
o
o
o
p-
p-
10
-
in
CD
0)
in
3-
01
p-
00
Z MEANIWITHOUT Q)
3-
ru
3-
in
ru
CO
cn
CD
co
ru
p-
3-
p-
ru
p-
cn
ru
ru
o
ru
STANDARD DEVIATION
m
nj
in
CO
CO
ru
o
o
p-
CD
3-
CO
-
cn
ru
p-
01
CO
CD
cn
CO
CD
Ul
.D
t—
a.
_i
o
z
<
t
X
3-
co
ru
in
00
ru
o
ru
ru
nj
cn
p-
ro
ru
ru
ru
CD
CD
ru
ru
STANDARD DEVIATION
O
1C
in
3-
2
ru
-
o
cn
CD
p-
CD
in
3-
ro
ru
—
Q
O
cr
Ul
Q.
O
h-
0
tr
o
p-
ru
CO
r>
o
i
3-
r~
o
cn
p-
o
cn
co
o
CD
CO
|
ffi
0
p-
)
o.
r-
0
r^
PO
r-
CD
0
<
t
O
>-
<
NJ
UJ
O
ooo
co cn o
ID UD tTi
co cn o
UD — r-
CD cn cn
OOO
*
CD cn cn
*
cn in tn
r^ cn o
00 —
ru co CD
co cn cn
ooo
* *
O l> IP
CD cn o
oo —
i UD —
CD 0
*
cn cn LP
co cn o
* m
CO i I
O
CD cn o
* *
r- o o
*
ru ru ID
r* cn cn
r- o o
*
LP t O
i_^
*
in i ru
»
CD I 01
P- Ol
in i— t—
I O O
C/l CJ CJ
r~i ~^ —
;.-
<
Nl O O
LJ -J _l
rr o _i
0 0
o ru
— ru
in in
0' OJ
o —
a o
* *
i cn
ru
i ro
ru
ir> ru
o ru
*
ro cn
0 0
i r-
— ru
0 —
*
in in
i in
in o
nl
ro r~
»
£ M
1— _l
0 0
cn
Ul
o o
_! LU
c/i a.
o o
o ru
P- 0
CO -
o ru
o m
co ru
o ru
ro ro
— ru
a ru
— cn
— ru
0 P-
ru o
CO 3-
0 —
0 0
ss
in 3-
o —
ru 3-
in ru
CD r-
o ro
B EPA BURTON CHAMP 1
T EPA BURTON HIVOL I
0 0
in -
O CO
— 01
f_^
_ CO
o r-
— CO
^^
ru r-
— r-
55
— o
0 CO
— o
— o
— r-
co ru
cn r-
_
r- -
— o
0 CD
0 00
S£
co cn
— o
0 3-
in o
CD ru
CO P-
o o
Q.
n
z cr
O UJ
D D
LU Ul
D UJ
o
1
1
in
cn
i
1
IP
ru
i
i
i
i
<
r-
ro
"
*
j-
IP
j-
o
cn
i_n
Q
UJ
UJ
_i
Ul
o
o
o
o
o
o
o
o
o
o
o
0
o
0
0
o
o
o
o
o
o
o
o
0
o
o
o
0
0
"— '
o
o
o
o
MEANtWITHOUT Q) 1
r-
-
CD
O
0
o
ru
o
CD
O
Ol
O
ru
ru
£
ru
o
(D
ru
ro
ru
CO
o
3-
ru
3-
ru
STANDARD DEVIATION 0
u.
u.
CJ
rr
o
>
in
r.
cr
a:
cr
UJ
UJ
CL
I
in
cr
o
cr
Ul
cr
UJ
REGRESSION VS Z INI
cn —
CD CD
0 0
ro p-
cn ru
' f^
0 0
0 0
ro in
CO P-
0 0
ro 3-
co p-
p- in
'
p- ru
ru P-
< in
i i
c/i in
O D
r z:
>- >-
< <
CO CD
D 13
Isl fx!
O Q
< <
CL 0-
LJ LU
o rr
ro co
CO CD
Ul Ul
0 0
CD cn
ru o
00 0
0 0
o o
J- J-
— r-
0 O
03 J-
O CO
01 IP
in —
3- 3-
f— t—
0 0
O O
c o
0 O
0 O
_l _J
_l _J
CO CO
_J _l
O _J
ro co
P- CO
0 O
M 00
j- m
o ru
0 —
o o
ro cn
ro co
u> o
0 —
ru CD
0 CD
0 f\J
'
r- ru
— ru
od
0 X
C/l
u
o o
0 0
_i rr
m a_
_! UJ
in o_
o ru
co in
0 0
3- ai
CD 01
in -
0 0
O O
ro j-
co j-
*—'
__
co —
in o
ro in
CD o
p- cn
o CD
i
B EPA BURTON CHAMP
T EPA BURTON HIVOL
3- in
p- co
Ul Ul
ooo
cn ro
3- o
0 3-
rt. rt
0 0
ooo
ai p-
— ro
f\J Cjl
— 0 O
CO 3-
o ro
o fn
oo cn
CO - 0
CL in
z: cn
< CL <
i i/> r
o t- ru
z cc o
O UJ LJ
y- —i --
fr _i _j
D H> LJ
co .E a
CL cr ^
U- Ul IP
DuO
-101-
-------
en
CO
>
ID'
i
in
8
cc
UJ
0-
tr
UJ
r1
o
CO
0
•ivaooa:
r
ui
—
_i
D
l/l
w
QCCO-IU'IOCDDU.O
inroroiDrur^cnr^iDOi
r^r^r~*r~r^(Dcncnr"iD
-r-riocoj-or-ruruco
iDiDiDtoiDr*~~~— ™r~in
i cJir^or^ror^jcnru
r-r~cor-r-oor-r-
r-cooajrocncoror-ro
inooruo i coinroro i
j-Tinm rur~r-in
r^ — ruroiuiDcncocn >
ru i r-j-jincoj-roro
ru rurururoj-j-ruro
i inruoocoiDruru i
rurorororoj-j-ru
ID J" i i i ro ro ru en r-
iniD — — ruiofo — — '
j-j-ininifij-r-r-in
' in § 5 £ 8 P £ S S
— njfuj-j-r-j-aioo i
j-j-j-j-j-air^r-j-
__..~.O i tDCDODCJiU)
JJ"J"T JiOlPJ-J"
U"J y i CD j~ o LO to t\t r*~
ronj i — o nj C7> r\j fu in
rofOKi&foaiiOinro —
i i o o o o r r i/i
QO x x i/i r
rrooooooccru
>• v z z or o
< • >- z z ct o
< < (ft O O UJ UJ
CD CD UJ t- t— _J ^
DDOOOOCCCt_/j
r-JMOOOODDDul
OO_J_J_IO:COCDI:O
ao.cococoQ.o!o-cci^
UJUJ— l—l-JUJUJUJUJl/1
o in
o ru
— o
0 3
o —
— o
o co
o ru
— o
o in
o —
— 0
O 0
o ro
— 0
0 01
o ru
— o
o —
o ro
— o
o ru
o ro
o in
o ru
— o
0 ID
— o
o en
o —
— 0
o in
0 —
— 0
0 J-
o ro
0 ID
o —
— 0
o r-
— o
o in
0 —
— o
o r~
O ro
— O
- O
ME AN (ALL SAMPLER?
STANDARD DEVI AT
i^, (** ifl j- Ol tO (D ~~" f*" f\J CT1
b-inoDQOCDCDrunjnjr-r"
LJOlCT>Ol(7)(TlC00)t7)CTlGD
O OOOOOOOOOO
g
>gSgSlgS§|2P
01 —
Orooooo — — — o —
or • •
UOOOOOOQOOO
uj o o, cn ro ru co ^ . r^ .
ooicnoiocnainjnjor^
mooo — oo — — — o
lS5SS££S£gS
uj o j" OT f) r~~ (\j oj o 10 j~
UJ l I I l I 1 — 1
z
i i o o o o £ z: i/i
oo liinz:
rrooooouccru
rsi
>• > z z o: o
in<< aioouiu
>ODOO Ld t— »— _i ^
DDOooooror_i_i
ZMMOOOO3DDUJ
in
UJCLO_CDCDCOCLCuCLCC>—
CcujuJ_j_j_JujuJujujin
o
UJ
QCCO_il/lOCODu.O
oroor<_>_iinocoDu.o
-102-
-------
r-
t-
01
00
-
j_
d
H-
C/l
Z
o
in
cr
•*f
0.
2:
o
CJ
tr
f
z
~
o
l/l
o
**
CD
in
f
—
ro
— .
01
.— 1
—
O
— .
cn
CD
f--~
to
in
j-
ro
ru
~-
Q
O
cr
LJ
Q_
l/l
_l
CO
LJ
cr
Q
00
01
o
in
ro
ro
0
ro
Ol
ro
r-
01
03
ro
01
ro
!
Ol
in
o
ro
ro
ro
in
Ol
ro
ro
i
LO
Q
r
>-
<
m
13
rsi
Q
LL
UJ
Q
tr
m
ro
„
CD
o
J-
co
ro
3
ro
cn
00
ro
r-
—
o
ro
03
3
OJ
ro
ro
j-
o
ro
r-
ro
00
Ol
0
ro
CO
i
CO
CJ
r
>-
<
03
rsi
Q
LL
LJ
cr
LJ _l
r- CD
j- in
r- in
CD r-
cn oj
3 in
U3 CD
in co
in 03
3 3
CD ro
in CD
r- o
ro in
in o
— Ol
i i
CD 3
It) ID
OD —
in LO
ID cn
10 CD
3- r-
in in
i r-
in
1 3-
in
1 01
m
t— »—
0 0
CJ CJ
Q Q
0 0
O 0
_l _!
CQ 03
_l -J
O -J
c/l
J-
m
ID
r-
_
ID
,
i
I
Ol
in
cn
— .
i
m
—
10
o
00
1
in
r-
r-
m
ro
CD
o
CJ
Q
0
0
03
_J
LO
0
Ol
OJ
r~
in
m
ro
cn
01
CD
01
„_
in
in
cn
cn
Ol
CO
3
Ol
3~
01
«
^
in
CD
ro
—
3
Ol
3-
3
Ol
r—
O
CJ
a
CO
LJ
CD
O
tr
u_
LJ
O
CD
03
-T
OJ
3
Ol
3-
3-
Ol
CO
ro
o
ro
j-
ro
—
• — .
o
3
ro
j-
03
„
ro
01
ro
01
ro
[ —
Ol
a.
i
o
z
o
or
03
CL
LJ
CD
Z>
01
3
ro
in
m
3
j.
ro
in
3
in
01
r.
ro
J-
o
ro
_*
3
m
3
00
3
O)
ro
cn
Ol
ID
ro
Ol
Q_
I
CJ
z
o
h—
tr
03
LL
LJ
z>
o
1
1
3
CD
r-
in
i
i
01
m
i
i
i
,
o
cn
o
CD
03
01
ro
OJ
Ul
LO
T:
01
0
LJ
UJ
0
r-
cn
o
r
ro
3
OJ
3
CD
CD
O
CD
o
3
00
ro
Oi
CO
3
in
— .
3-
ro
— .
m
CD
3-
CD
in
o
3
ro
3
o
3
CO
ro
a
t—
33
0
i
t—
3
• — •
Z
LJ
rsj
—
Ol
ro
ro
in
03
J-
Ol
01
03
r-
ro
—
CO
—
ro
—
ru
i^
—
Ol
in
z
o
^
- — -
LJ
Cl
0
cr
o
^
r~
C/l
cn
ro
cn
cn
^
r^
0
3
CO
ro
ro
3
01
in
in
• — .
r-
ro
CO
3
Ol
3
LO
in
_
3
O
3-
Ol
ro
ro
ro
13
0
Q
' — •
Z
LJ
X
ro
—
„
—
o
m
^
—
O
—
r-
—
o
ro
—
— -
01
Ol
—
Ol
01
OJ
—
ro
— •
Ol
3
-z.
0
t—
"-•
LJ
Q
Q
cr
C3
z
t—
C/l
>
CJ
o
o
**
ID
Jl
—
3
~
ro
~-
at
—
— ,
—
0
—
01
03
[^
CO
in
3
ro
OJ
^~
o
o
or
LJ
0.
rsi
O
» —
O
h-
_
<;
03
3 0
rsi o
Q _l
LL CO
LJ _J
cr cj
3 t~~
00 ro
— Oj
o o
o CD
ro ro
- -
_l CO
ro 00
ro 01
_- _
3- CD
OJ O
*~" "~
CD o
03 0
O —
•
m i
CD
—
in i
Ol
„
O 1
in
—
Ol 3
O —
— • — -
Ol Oi
Ol Ol
_- _
*
1 1
in r-
Ol 3
_ »
J- ro
ro ro
— —
Ol 00
— ro
-j. |
3-
„
O! ro
ro r~
3 ~~~
ro 3
— «
3- CD
3- r-
„ „
o o
CJ CJ
Q O
0 0
0 0
_J _J
(33 UJ
_l _l
_l LO
cn r-
ro j-
in —
0 0
o —
- CO
~ o
O 03
03 ro
CD -
0 —
*
CD (7)
Ol CD
o o
•
Ol o
01 r-
OJ a
Ol o
r^ 10
0 0
Ol o
r- o
o —
• •
— - __
01 r-
— o
•
ID ro
o r-
Ol O
r- o
03 f-
— o
*
r- 01
ro oo
— o
•
— in
00 03
0 0
•
nj in
3 Ol
o o
cn ro
cn oo
0 0
•
— [^
cn r~
0 0
•
3- 3-
Oi r-
a o
in 03
o 01
— o
CD CD
CD r-
0 0
*
t- 0-
o ^
o <
55
z
c/l O
LJ 1—
o cr
O D
a: CD
U- LL
Ld LJ
O 03
LO
ru
Oi
o
„
CO
o
D
CO
cn
o
03
CO
0
in
r-
o
in
03
o
OD
—
„
01
in
o
o
03
O
CD
Ol
o
Ol
CD
O
Ol
o
cn
cn
o
OJ
03
o
cn
01
o
*
CD
CD
O
O
cn
o
10
CD
o
CL
j:
i
CJ
-^
0
t—
cc
cn
o_
LJ
D
C
ru
LO
0
in
-
o
,
i
0
J-
-
J.
f
—
1
1
r-
ru
^~
i
i
i
i
(
CO
ru
01
CD
ro
LO
j-
ru
ro
to
o
in
in
ru
o
LJ
_J
Lu
Q
1
in
G
o
o
_
o
o
—
o
o
•""
o
0
-
o
o
—
o
o
_-
a
o
—
o
o
— 1
o
o
,-
o
°
CJ
o
— .
0
0
—
o
o
o
o
_,
o
o
o
o
— -
o
o
„
o
J
ID
o
X
t—
3
~z_
LJ
CO
ro
o
l£>
ru
o
o
ru
o
in
U~l
o
CD
PO
O
—
aj
o
j-
ro
o
r^
j-
o
0
in
O
—
Ol
o
to
Ol
o
in
ro
o
ru
OJ
o
cn
ru
o
cn
ro
o
o
ro
a
ro
j-
o
z:
0
t—
i — .
UJ
o
0
a:
o
z
»~
in
u.
u.
LJ
o
cc
o
o
u
Q
1/1
cc
cc
o
cc
a;
UJ
LJ
a.
o
LO
cc
o
cc
tr
LJ
i —
a.
LJ
CJ
tr
LJ
Z
Nl
LO
Z
o
LO
l/l
UJ
cr
o
U)
in
01
03
o
3
r-
03
3
in
Ol
o
Ol
CD
00
o
p^
Ol
ro
in
r-
00
Ol
ro
<
C/l
Q
^
t£
03
D
rsi
Q
0_
LJ
Q
„
J-
03
o
3
CO
ro
in
in
Ol
o
03
OJ
f^
o
J-
03
in
m
ro
03
j-
LO
LO
Q
•^
<£
CD
Nl
O
CL
LJ
CC
o
CD
r-
o
o
OJ
ru
cn
10
OJ
01
o
OJ
00
o
f--
cn
j-
o
o
— .
to
o
t—
o
o
Q
O
0
j
03
C_)
o ro
03 —
r~ CO
o o
in o
3- _r
— 3
03 0
in to
01 in
— Ol
o o
ro oi
00 0
o —
ro —
Ol OJ
r- o
03 Ol
01 —
J- 3-
01 in
r^ oi
H- >r—
o o
CJ O
Q Q
O O
O 0
_l _J
CD 5
_j LO
ro
CD
in
CD
nj
03
01
tn
ai
03
03
in
0
_
CD
in
—
in
OJ
01
to
nj
jr
. — .
UD
p^
—
1
O
CJ
Q
LO
LJ
Q
0
tr
a.
LJ
O
Ol
Ol
r-
o
Ol
—
LO
Ol
3
O
O)
CD
t-
O
f^
3
ro
CD
ID
Ol
3-
J-
a.
T:
i
CJ
^y
O
1—
tr
03
0.
LJ
CD
Ol
ro
03
o o
J-
in
CD o
in
o o
CD
o
01
0 0
LO
O
o
r- o
Ol
o
ro
ro o
0. l/l
i; i/i
i x:
CJ Ol
Z Q
O LJ
t— -.
tr _i
D LJ
03 O
CL P
LJ LO
U O
-103-
-------
p-
p-
01
OD
™
v
<
11
UD
I
CO
8
cc
CJ
Q.
..
cc
LJ
h-
U
U
§
o
•^
\f)
<
5
o
z
z
ro
i
_i
2
0
U.
o
s
h-
CO
z
o
LO
cc
Q.
r
o
o
cc
LJ
Z
—
8
o
UD
in
r
ro
ru
_
—
o
O)
OD
P-
co
in
j.
ro
ru
_^
Q
O
CC
LJ
a
CO
t—
1
^5
CO
LJ
CC
rr i-
o in
ru co
ru
o j-
in ro
r- co
0 3
ro o
in co
o ro
p- in
0 p-
o cn
ro ro
in co
0 0
ro ro
p- o
a —
p- ro
3- in
O P"
ID P-
3- P-
o ro
3- cn
— m
0 0
co in
— in
ru —
o r
in —
ru o
•"• —•
o in
- UD
CD ro
— ru
o cn
_ 00
r- o
0 p-
UD CD
ro —
— ru
o in
in ru
cc cn
o r~
p- co
ru ru
- ru
To1
CO >
D —
z: i
>- z
< 0
CD t-
D CC
NJ D
Q 00
*t ^
Q. CL
LJ LJ
CC 1-
< LJ
O O
P- cn
in cn
ro ro
0 O
cn UD
ru
0 O
in cn
ro ro
0 O
oo in
cn — "
ru
0 O
ro 3-
j- O
ru ru
o o
P- in
o CO
0 O
p- r-
ru cn
1 O
in
—
0 O
ru cn
- in
j- in
0 O
ro o
ru ro
o o
ru ro
3- t
0 O
3- 3-
p- ro
ru ro
0 0
cn ru
ru o
ro ro
0 1
ru
ro
0 0
OD 3-
f ru
ru ru
O 1
—
ID
_J
O
> 0.
— LO
I H-
CC CC
LJ LJ
Z _/
Z _l
i- r
i ^_
z cc
CO LJ
< LJ
ro
p-
cn
o
00
00
CO 3~
ru -
cn ro
?s
—
cn nj
— P-.
- ru
ru —
OD ru
ro ru
rxi iD
0 CO
ID ~"
T fTl
J U'
o 00
p- p-
~ T
o in
co m
o 3-
ru -
UD m
00 t
— o
UD cn
00 ro
ru ru
in ru
UD P-
ro 00
ru
r ru
ro in
P- CD
ru —
3- ro
ID 3-
in co
ru
P- ID
ru —
o ro
ru —
ru —
ru co
ru co
ru
3 ID
ru in
CO OD
co r-
— o
r- in
z
~ o
LO —
CC t-
LJ rf
Q. >
*t O
CO
_J <
< a
- Z
z <
LJ CO
NJ
oo cn
T cn
ru —
in r
— p-
3- 0
in ru
— 00
ro 3-
ru —
3- 0
cn in
ru r-
o cn
r- ru
r- co
p- ru
- CO
ro 3-
r~ • '
in co
ru ru
j- in
CO O
0 r
3 00
ro UD
j- in
ro ru
p- p-
ru in
f 0
ru —
OD CO
nj —
ru cn
ro —
in ru
CO 3-
ru
CO O
CD o
3- ru
ru —
3- 0
in co
CO CO
ru
ru r-
oo ro
CO CO
cn in
ro ID
cn 3-
z
o
K-
«J
• — .
LJ
- Q
LJ
• a
< CL
i- a
- z
z <
< t—
LJ CO
r
X
o
o
CO
o
co
in
3-
ro
ru
_
—
o
_
cn
00
p-
UD
in
t
ro
ru
_
o
o
a:
Cu
a_
0
o
^-
i: i
>- z
< o
00 t-
D CC
rsi D
Q 00
*< *t
0- 0-
LJ LJ
CC 1-
P- UD
ru —
ro j-
ro ru
o o
j- ru
ru in
< LJ
cn UD
to in
in co
CO p-
0 —
cn —
3- in
— --
cn 3-
i^ r-
o —
P- 0
co 3-
„ „
— 00
cn in
o —
•
o ru
ru cn
— o
i r-
in
—
3- in
j- cn
o ru
cn in
o —
r~ o
in UD
~ ~~
p- o
o ro
— i ^~
•
ru cn
co jr
0 1
3-
ro co
in ro
:
3- '
cn
o
*
0
> Cu
— CO
I i-
a: a:
LJ LJ
Z _!
Z _J
< 3
_J H-
z o:
00 LJ
< LJ
o ru
o in
— o
o cn
o in
— 0
o o
O CO
— 0
o o
O CO
— o
0 0
o m
— o
o ro
O CO
— o
o —
o j-
— 0
o ro
0 3
— o
o ru
o in
— o
O 3-
0 00
— 0
0 p-
o ro
— o
o OD
0 UD
— 0
o in
o ru
— o
o m
O CO
— 0
o o
0 3-
— 0
o ro
o in
— o
0 0
o ro
— o
Z
- O
cn —
rr i-
LJ <
CL >
£ LJ
< a
cn
Q
_l CC
< Q
— z
z <
< 1-
LJ CO
r
U- P-
U. 3-
LJ CO
O •
0 0
S
o
• in
> in
o
• CO
co ro
n 3-
cc
cc -
cc •
CJ 0
LJ —
Q- CO
o in
CO O
CC ro
O —
o?^
LJ cn
3"
ro
f— cn
CL ru
LJ 3-
CC "
LJ 3-
Z ~
CO
1
CO
D
r
NJ
CO <
> CO
D
Z NJ
O Q
CO
CO <
LJ CL
rr LJ
o
LJ
cc rr
oo 3- ru
— co cn
o o o
— ru co
r- co p-
o ru cn
cn in co
ru oo ro
co in 3-
in ru cn
in UD P-
ru ru —
000
ru ro co
3- ru 3-
— UD P-
o — —
3- — 3-
UD ru ro
— ru ro
3- ru 3-
o ro in
in in ro
UD CO —
cn — ru
o — —
o O OD
— 00 P-
cn co ro
i i
o o
> > 0.
CO
III-
z rr cc
O LJ LJ
1- Z _)
cc z _i
CD i- r
< _i i—
CL Z CC
LJ CD LJ
»— < LJ
-104-
-------
p-
O)
CO
>-
<
10
cn
8
LJ
0.
CE
*—
LJ
T.
(_>
S/CUB
<
5
o
z
z
o
2
_l
_)
£
in
u.
o
>-
Q
3
1—
in
Z
O
in
cc
RCOMP/
h~
Z
Q
l/l
O
>
<
ID
in
3-
ro
nj
o
01
CD
p-
10
in
3-
ro
ru
-
RESULTS PERIOD
CC OD
O (O
(O in
o in
in o
o —
in 3-
o ru
in 3-
o in
in 3-
v m
o —
in P-
0 0
in o
o cn
in 01
o 01
5 M
0 P-
in P-
o in
in —
O (\J
m ID
0 rO
in r-
0 f
in CD
a in
in in
o a
0 3-
Ul 0-
sl
E 0
>- z
< o
CD I-
3§
Q CO
< <
0- 0.
LJ LU
cr m
D U.
p- a
m m
nj
in o
— P-
— o
— ru
txj
i
en o
ru p-
in o
ro CO
3- ro
CO O
— 3-
ID 0
en r-
ID o
in o
in o
in o
to
i
0 0
ID f
— O
0 —
ru
en o
j- in
i
CO 0
fO CO
nj
nj o
ID p-
— M
i
0 0
fO CO
nj
CO 0
p- ID
Q.
£
< Q.
I LO
U CL
z cc
O LJ
t~ 1
CE _l
D 13
CO £
< t-
0. CE
LJ LJ
3 U.
ID
P-
in
0
3- a
ru ro
0 O
fO P-
— ru
CO CO
3- 01
— O
P- f-
CO —
p- in
LD —
p- ru
in o
in in
co ro
— p-
— ro
3- 3-
— OJ
ru O>
01 3-
0) r-
0 —
3- ID
in —
r^ ro
a) ro
01 01
ro 01
ru —
co ro
co ro
ID P-
ru r
2 ME AN (WITHOUT R) \i
STANDARD DEVIATION 6
cn co
o nj
ro in
j- o
ID CO
— ro
o ro
o m
oo m
O ID
cn 10
01 —
ru co
£2
co ro
ru 3-
OD r-
OJ O
ai nj
0 J-
o ID
to -
o ru
0 p-
o m
cn ~
M ^
ID 3-
ID 3-
ru —
in p-
o ro
cn 3-
ID CO
X MEAN(U.F) 1L
STANDARO DEVIATION t
>
u
Q
in
o
>
<
a.
r-J D
Q CO
< <
Q. CL
Ld LJ
CC CD
f nj
ro ru
o ru
ru 3-
0 0
CO ID
m co
o —
D U
ru p-
ro ro
o ru
0 ID
in co
o —
ID 3-
in o
o ru
o x
r- in
0 ~
in in
in r-
o —
3- O
in LD
o ~
in ru
0 0
*
p- m
p- ai
o —
«
ro ro
ru in
o ru
ru co
in o
o ru
in ID
3- nj
o ru
3- ru
p- in
o —
3- cn
in —
o ru
r- in
in co
o ~
3- ru
o ru
0 0
CO ID
o —
0.
i&
o a:
z rr
O LJ
i^
co s:
< i-
Q. cr
LJ U
3 L.
o m
o r-
— 0
o co
0 —
o in
0 P-
— o
0 0
0 01
— a
0 0
o in
— o
o ID
o ID
— 0
0 3-
o m
— o
0 p-
o o
— 0
0 3-
o fXJ
— o
o ro
o ro
0 3-
o cn
— o
0 O
o —
o in
0 3-
— o
0 f
O D
0 3-
0 p-
— 0
o in
0 0
o ro
o in
MEANIWITHOUT R) 1
STANDARD DEVIATION 0
u_
Li-
o •
<_> o
g
>
LJ
O •
tn
*:
tr
o
o:
tr •
LJ O
LJ
0-
_l •
l/l O
CE
0
CE
CC •
CL
LJ
O •
CE 0
LJ
REGRESSION VS Z INI
R EPA DZUBAY MDS-S
ID ro ID
r- cn o
000
ru ru ID
o o in
3- o ro
ru o —
aj ru 3-
in 3- co
01
000
— p- OJ
ru cn co
o o ru
— 3-00
ro cn —
CO CO CO
p- in o
— — ro
cn — o
P- ID f
ro m cn
— ID P-
B EPA BURTON CHAMP e
U EPA BURTON CHAMP
F ERT MULLER RSP -I
-105-
-------
f-~
r-
cn
CO
i
~
^
«£
^"
ID
|
— .
cn
Q
o
cc
Ul
CL
CC
Ul
UJ
i:
(_j
> — .
CO
D
o
Ul
^
cc
o
o
2
2
fl
O
2
Ul
o
cc
-I
>_
Q
f-
in
2
in
cc
0.
£
O
O
cc
UJ
t—
z
—
Q
in
o
ID
in
j-
r*i
._
ru
^^
0
cn
CD
(^
10
in
j_
M
ru
«
8
a:
UJ
Q.
in
H-
_l
=>
in
UJ
cc
cc
0
UD
ru
0
in
—
o
CO
cn
o
r^
0
—
o
r*i
in
0
I—
o
r-
j-
a
ID
o
j-
— .
o
CO
ru
o
in
v«
o
CD
0
O
10
ro
—
o
g
O
CO
1
(/I
^
CD
Nt
Q
0-
CC
0
Ol
00
o
0 0
to
ru
0 0
in
— '
CD 0
CO
01
o o
r-
0
—
o o
ro
in
o o
M
r-
0 0
r-
o o
10
j-
o o
J-
„
o o
CO
r\j
o o
in
ru
0 0
CO
O CD
O O
ID
—
0 0
in
CO
o o
CO
— o
in —
CC l-
Ul <
CL >
in
Q
_l CC
< Q
2 <
< t—
UJ in
NJ
1 O
1 O
1 O
I O
1 O
1 0
1 O
1 O
1 O
i a
1 O
1 0
1 CD
1 O
1 0
1 O
z
o
£_
«^
1— r
;>
UJ
D
Q
CC
- Q
Z <
< t-
UJ (fl
X
a
in
o
ID
~~
in
j-
—
M
—
OJ
~™
_
0
—
cn
CD
f^
ID
in
3-
ru
—
Q
O
tr
LJ
a.
NJ
o
t—
0
t—
a:
0
o
o
cr
0
0
*
o
0
*
o
o
—
o
o
—
o
o
*
o
0
—
*
0
o
_
*
0
0
— •
*
0
0
^™
*
0
o
_—
*
0
o
—,
*
o
o
*
o
o
_!
0
o
~"
*
o
o
«
*
0
o
~—
*
01
(/)
0
X.
>-
CO
Nl
Q
0-
Ld
cc
o
o
— 0
o
o
— 0
o
o
— 0
0
CD
— o
0
o
— o
a
0
— o
0
o
— 0
o
CD
— o
o
o
— 0
o
a
— CD
o
o
— o
o
0
— o
0
o
— 0
o
o
— o
o
0
— o
0
o
— 0
o
o
— o
2
— o
Ou t~
Ul <
CL >
11 Ul
< Q
in
Q
_l CC
< a
- 2
•< V-
LJ l/l
j:
U. CD
U. 0
UJ O
o •
CJ —
cc
o
o
u
o •
• o
in
i:
cc
cc
o
cc
cc •
U) 0
LJ O
n o
O o
in -
cc
o
cc
cc •
Ul CD
t—
a.
Ul
o •
CC 0
UJ
t—
2
—
in
in
Q
M
>-
en <
> m
Z M
o o
in
in <
Ul CL
cc u
o
u
cc cc
-106-
-------
o o cn in —
r-
01
x
x:
o
o
cc
LJ
CL
cc
LJ
UJ
*-
(_j
^-
X
D
^
l/l
CC
Q
^
z
..
3-
o
in
_j
t—
o
t—
u_
o
V
o
D
H-
in
z
o
l/l
or
|
2L
O
u
cc
Ul
z
~
Q
in
o
ID
in
3-
ro
(XI
-
o
Ol
X
r-
ID
in
j.
ro
ru
_
Q
0
or
LJ
CL
in
K-
1
^3)
in
LJ
cc
cc Q. t-
000
ro o o
01 cn x
r- o x
ru ro ru
0 0 O
0 O 0
ID rji r-
x 3- in
ru ro ru
000
o o in
cn ro 3-
ru ro ro
000
X o o
o 01 ru
3- o in
ru ro ru
i
o o o
0 0 O
in x ru
ro r~ x
o o o
in o o
r- r- in
3- CD ID
O 0 0
r- o cn
000
o o 01
o r- m
— r- 3-
000
O O ID
o cn 01
3- r~ in
ru ru ru
o o o
ro o ro
ro ru ro
ru tn ro
ru ru ru
0 O O
3- o x
3- ro 3-
— r~ in
ru ru ru
o o o
ru o 3-
x — in
r~ ru o
— ru ru
000
o o in
— in in
o o o
r- o x
ID 3- 0
— ID ro
_« _« _~
000
ro o ID
.XI 0 |Vi
o 01 in
000
— o in
Ol ID ID
ru r- ro
— — _
in _j _i
i 0 0
in > >
o
r i i
> z
< in o
X U 1—
D Q CC
MOD
Q CC X
CL
— in
i i-
cc cc
U LJ
Z _l
< D
t- £
-J t™
Z CC
CD ui
< LJ
o
ro
3-
ro
01
ID in
28
x —
ru
O ID
ID r-
(XI —
x 3-
ru
o r~
O) ID
ru ro
o ro
ro
ID o
in in
r- 01
in ru
(XI
0 X
X O)
ID -
?*
in ru
0 0
ru 10
3- ro
ID ru
ro 3-
r~ ru
ru r-
ro ro
__
ro o
£8
m —
ru
ru r
r- o
r- x
ru -
ru
3" X
x x
x 3-
ro ru
ru
ru o
— o
r- in
x ru
o —
O) (XI
3" cn
ro —
ro ru
— ru
— ro
ro ru
_
x in
— Ol
3- ID
ro ro
in r-
— 3-
3- —
3- ru
^
— O
in —
CC K-
LJ <
Q- >
i: uj
< Q
in
D
_i or
< Q
Z <
< 1-
u in
r
tsl
o ro
in r-
x ru
ru
in m
r- —
— x
x 3-
ru
in x
o ro
ro
in cn
r- ru
— ru
ID ro
ru
o uo
S -
0 —
in 01
ID 3-
in ru
x r-
o —
01 3-
ID ru
o r~
ro ID
o —
3- 3-
O (XI
ro r-
01 ro
10 —
ru
ro ro
§lD
0
ru ru
(XI
in —
O) 0
3- 3-
3- ru
ru
in 01
ro ru
01 CO
x ru
x 3-
X 01
3- —
ro x
Ol OD
in in
ro ru
_«
in x
~ ro
ru r~
3- ro
r- in
— ru
01 ro
3- ru
z
o
h-
^
— • ^>
UJ Ul
- o
- Q
i- CC
O. Q
- Z
z <
< 1-
ui in
r
X
Q
l/l
0
ID
m
ro
ru
-
o
Ol
X
[-^
ID
in
3"
ro
ru
_*
O
o
cc
LJ
CL
0
t—
o
~^
b—
cc
cn
o
o
o
cn
0
cc
cn
cn
o
o
%
3"
Ol
o
ro
O)
o
Sj
o
o
0
ro
X
0
ro
cn
o
X
cn
o
o
Ol
o
m
Ol
o
ru
x
o
01
x
0
ID
r-
o
o
Ol
o
in
i
in
Q
r
V
<
^
Q
^
Q.
Ul
CC
ru 01
— o
o o
— r-
(XI o
CL t-
o ro
— o
ru 01
— 0
O) ro
o —
0 X
ru Oi
— 0
*
o ro
— f-
(XI O
— —
o cn
in ru
ro ru
ro —
X 0
0 0
•
— ru
— o
— «
*
3- r~
— o
_ «
X o
__ _—
^~ — ~
ID ro
— —
in o
ru o
_^ _^
•
ru in
3- —
ru in
ru oi
— o
•
_j _j
O O
> >
I I
z
in o
LJ I-
CC X
^ ^
CL CL
Ul Ul
CL »-
x o
o —
o o
X (XI
x cn
o o
< u
ro ID
Oi Ol
0 0
in o
X O
o —
in r-
x cn
o o
ru r-
Ol O!
o o
• *
ru ro
01 o
o —
3- ro
X Ol
0 0
cn ro
r- r-
0 0
i ru
r-
o
i i
cn o
x o
o —
«
- x
cn 01
o o
* •
x cn
x x
o o
— x
cn O)
o o
*
ID i
X
0
ID (XI
X X
0 0
j> [
Ol
0
*
_,
o
> a
— in
i i-
tr cc
UJ Ul
z _i
Z _i
< D
i- i:
i ^_
z cc
X LJ
< u
0 10
o —
— o
o r-
0 CD
— o
o in
0 —
— o
o —
o —
— o
o —
o —
— o
o ru
o —
— 0
0 3-
o —
— o
o r~
o ro
— o
0 X
o ru
— o
0 X
o o
— o
o x
o o
— o
0 O
o —
— 0
o ro
o —
— 0
0 3-
O —
— 0
0 X
o •—
— o
0 X
o ru
— o
o in
o —
— o
z
- o
in —
CC H-
U) <
0- >
i: ui
< Q
in
o
_l CC
< Q
- z
z <
< r-
ui in
u. 3- ro
u. 01 x
LJ Ol Ol
o • •
O 0 0
cc
o
o
3- —
> ru ro
UJ — • 3"
Q • •
ru x
in r- o
r x 3-
cc —
cc ru —
O ro in
cc o o
cc • •
LJ O O
u cn cn
0. ID -
o o o
i
in
__
O (XI X
cc in —
or • •
uj — ru
m in
ID o
1- 3- (XI
CL x ru
LJ Ol ~~
O
or ID ru
LJ r- r-
I- X 0
z ru ro
in _j
1 O
in >
Q —
M
in < in
> X LJ
D Q
Z M O
o o cc
in
in < <
LJ CL a.
CC LJ Ul
O
Ul
cc cc a.
x 3- —
r~ 01 cn
01 cn cn
000
ro — —
3- O) 3-
0) 0 -
o ro 01
— 3- Ol
in r- o
in 01 ro
in ru 3-
o o o
000
r- ro 3-
in cn ID
cn x o
0 O —
ro
ID O O
r~ o ru
x r- 3-
ru x o
— in cn
ID 0 0
x ru —
X (^ X
3- ru ru
in in ID
— 7 ru
i
_j _j
o o
> > CL
— — in
i I >-
z or or
O LJ LJ
l- Z _l
or z _J
D < D
x h- z:
^ ) K-
CL z cr
UJ X LJ
1- < LJ
-107-
-------
r-
0)
CD
1
10
I
8
bJ
CC.
u
u
5
o
NOGRAI-
o
in
in
u.
o
1
t—
§
in
or
o
t—
o
in
0
ID
in
3-
ro
ru
-
0
cn
CO
r-
ID
in
3-
ro
ru
-
RESULTS PERIOD
CC O
00 0
ID ro
ID t-
ru ru
r- o
O (D
r- in
ru ru
8! '
in
CO
ru
oo o
ID o
ro f*>
ru ru
ID o
oo ru
ru in
ID o
t- o
ro —
cn o
cn in
ro o
on —
0 O
in o
ru ro
ru ru
CO o
in r~
ru ru
in o
— r-
0 —
ru ru
cn o
ro ro
CO o
r- o
— 3-
in o
O CO
ro o
r- ru
cn —
cn o
r- o
ru —
1/1 »-
1 0
l/l CJ
o —
£ Q
< in
CO Ld
rsi o
Q cc
Q. Q.
Ld Ld
CC. O
CD D L_
ID 00 0
3" CD 0
oo oo in
ru ru ru
r- in o
ro cn o
cn cn 3-
ru ru ru
ro oo o
— ro o
r- 3- 3-
ru ru in
ro ro ru
000
o in o
CO 00 —
ru ru ru
o — o
o CD o
J- r- ro
OJ J- 0
j- ro o
UD in nj
r^ t~- o
CD CD 0
r- to o
r- ru a
— ro r-
lf) CD o
ru ru —
ro ru o
ro ID o
ru r- ru
ru ru ru
ro cn o
3- in o
ID r- 3-
ru ru ru
ID ru o
in ru o
ru ru r-
ru ru —
ID in o
— — o
ro 3- o
cn cn o
ro ru o
— 0 0
3- o o
3- ro cn
in ro o
oo 3- o
3-3-3-
B EPA BURTON CHAMP
U EPA BURTON CHAMP
F ERT MULLER RSP
ft
%
cn
cn
00
co o
3- o
ru
o ro
CD CO
r- ru
ru
CO ID
oo t-
cn ro
ru
ru
r- f
in o
o in
m o
ro ru
in ~~
3- cn
o to
0 —
00 ID
ru in
ru
in in
ID ID
ro ru
ru
r- in
cn o
3- ru
ru
cn in
ru ro
cn o
ru ro
o —
— ro
in ru
— ru
r- 3-
ru 3-
— (U.
— 0
3- 3-
ro —
Z ME AN (ALL SAMPLERS)
STANDARD DEVIATION
ro ro
3- 3-
r- ru
ru
in ro
CO 0
ID ru
ru
cn r-
— r-
01 cn
00 3-
ru
ro ID
co cn
3- ro
ru
o in
o ru
in —
— CO
— 0
ro ru
cn 3-
r- o
ru cn
o aj
to ro
— ID
ru
r~ r-
ro ro
ru
ID in
r> cn
3- ru
(U
r~ o
— o
3- ro
ru 00
r-- ru
o in
ro 3-
o ru
0 —
3- o
3- in
— 3-
ro ~
K MEAN(O.U.F)
STANDARD DEVIATION
Q
1/1
O
ID
in
3-
ro
ru
-
o
cn
CD
r-
ID
in
3-
ro
ru
-
Q
O
CC.
Ld
CL
rsj
O
O
CC
cn in — r- —
0 — — 0 —
0 0 0 0 O
ro cn 3- ID CD
cn co — — CD
0 0 — — 0
CC O 00 D U.
00 — 3" ID —
cn o o o cn
0 — — — 0
*
— ru oo oo —
o cn o o cn
— 0 — — 0
ID i o cn in
cn — o co
0 — — 0
3- ro ru tn ID
cn cn — — oo
o o — — o
r> oo oo r- —
co o cn — cn
o _ o — o
cn 3- cn — r-
cn oo — — OD
O 0 — — 0
o 3- in in r~
r- oo 3- ru r-
*
r- cn — — ru
cn cn — • ru t~-
o o — — o
*
— ru co co o
— — — — o
*
3- — cn cn co
cn cn cn — cn
o o o — o
* *
ID — cn ru ru
co cn o — o
0 0 — — —
*
— ID co r^ co
cn oo — — OD
ro ro 3- ro ID
o j- aj ro cn
— 0 — — 0
r- 3- ro in —
0 0 — — 0
ro cn ru 3- nj
co cn ru — oo
O 0 — — 0
— in o r- r-
cn oo — o o
0 0 — — —
R EPA DZUBAY MDS-S
0 EPA RODES DICOT
B EPA BURTON CHAMP
U EPA BURTON CHAMP
F ERT MULLER RSP
0 ID
0 —
— 0
O ID
o o
— o
o OD
O 0
— 0
o ru
o —
— 0
o ro
0 —
— 0
o ru
o —
— o
o in
o —
— o
o ro
o ro
o oo
0 —
— o
0 3"
o nj
— o
0 —
0 —
— o
o —
o —
— 0
O ID
o —
o in
o ro
— o
o cn
— 0
0 CO
o ~
— 0
0 —
o —
— 0
ME AN (ALL SAMPLERS)
STANDARD DEVIATION
L_
Lu
(J
CC
0
Ld
in
or
o
cc
CC.
Ld
Ld
CL
O
in
cc
o
CL
Ld
o
CC
Ld
•z.
REGRESSION VS 2
— o
cn r-
0 0
co ru
co cn
— 00
ID OD
0 OD
ro r~
0 0
o o
cn ro
cn o
o —
cn co
ID —
i\j r^
ru ro
r- ro
cn ru
ID 3-
cn ru
ru —
i ru
in i-
I 0
in o
D —
r Q
< in
CO Ld
D Q
M O
Q CC.
CL CL
Ld Ld
CC O
r~ ro ru
co cn in
o a o
in in r-
3- o m
oo 3- ru
in cn ro
ru cn ru
— ru
3- ro r-
000
o o o
co cn —
o in cn
0 0 00
— — o
in ID o
oo 3- ru
oo — in
3- r~ o
oo 10 in
in cn 10
— — cn
UD ID Cn
in ID in
B EPA BURTON CHAMP
U EPA BURTON CHAMP
F ERT MULLER RSP
-108-
-------
r-
r-
cn
CO
t
~
v
"4
y
CO
—
cn
o
cc
U)
Q.
..
CC
LJ
UJ
O
CO
D
cn
z
o
z
z
..
5
in
LJ
o
fy
^
—1
U.
O
Q
t—
cn
-z.
o
in
cc
a.
§
0
cc
LJ
t—
Z
—
a
o
ID
in
3-
ro
ru
_ .
_~
o
cn
CO
^
ID
in
f
ro
OJ
— *
Q
0
1—1
CC
LJ
0.
in
t—
z>
in
LJ
CC
ce o
0 0
LO ro
O 0
— —
00 0
co r-
co o
o o
ru
cn o
o co
ID CO
0 0
r> —
in PO
—
in o
r- 3-
cn 3-
r- o
ID 3-
CO CO
LO O
0 0
in ru
r> o
— o
0 (D
— —
— o
r- r-
in 10
r- o
ru cn
IT) U)
ru o
CO I-
ID O
—
cn o
ru in
CD ro
in o
r~ PO
co in
ro o
if) o
3- ru
nj o
ro in
in 3-
in >-
l O
LO U
Ti Q
< in
§ Q
rsi o
Q CC
^ ^
Q- Q.
LJ LJ
CC O
ru
0
3"
CD
3i
in —
3- ru
0
—
Cl 01
r- ru
01 —
in 3-
— _
in cn
J" O)
ru co
o ro
3- ru
a> in
co cn
o nj
ru ro
3- cn
ru oo
M
M ID
in —
ro r,
cn ru
o —
ro 3-
- r-
ru r-
— r-
cn in
o —
ID -
ID 3-
r- t-
CO ru
0 0
cn —
cn in
M 3-
0 J-
r~ ru
I — Cn
ru r-
— CO
cn in
3-
Z
~ O
in —
cc »-
LJ <
Q. >
r LJ
< Q
cn
o
-I CC
< o
— z
z <
< t-
LJ in
T"
rsi
o o
0
~
O 0
r>
0
o o
w
ru
0 O
CO
CO
o o
PO
—
o o
3"
0 0
f^
0 O
o
ru
0 0
o
CO
o o
r-
co
0 0
cn
(D
o o
r-
0
r~
0 0
in
ro
0 0
m
o o
0
ru
0 0
in
3-
Z
O
«f
;>
UJ
Q
a
cc
O Q
- z
z <
< H-
uj cn
•c.
X
o
a
o
Id
in
3"
ro
ru
—
__
0
en
CO
r-
co
in
r
ro
ru
—
O
0
cc
LJ
Q.
Nl
O
H-
O
t—
CC
cn CD
ro ru
ro ro
ro ro
0 0
x co
CD —
O —
CC O
— 01
o cn
— o
• *
— cn
cn o
o —
*
in in
ro CO
o —
cn -
3- in
0 —
— cn
ID ro
0 —
— cn
CO —
o —
•
ro r-
m 3-
o —
ro r-
3- in
— 0
*
Co ru
r- ru
o —
— cn
in 3-
o —
r~ ro
CO —
0 —
*
co ru
r~ ru
0 —
*
3- ID
(D ro
o —
*
in in
ru r>
— 0
*
cn —
ro ID
— 0
*
co ru
o cn
— 0
*
cn i-
I 0
in o
r o
< in
m LJ
D Q
NJ O
o a:
— 0
O (D
o in
— 0
o r~
o ru
— o
0 ID
0 CO
— o
o —
O ID
— 0
o ru
o ro
— o
o cn
o ID
— 0
o cn
0 —
— 0
o —
o ro
— 0
o ru
o in
— o
o m
o ro
— o
CD m
o in
— o
o ru
o —
— 0
z
- 0
in —
CC 1-
LJ <
_J *—
Q. >
£ LJ
< D
in
Q
_l CC
< Q
- Z
z <
uj in
z
u. ru (D
LJ ~r cn
o • •
000
cc
o
o
• CO ID
> on 01
LJ CO ID
. ~— —
cn co cn
^
Sin in
f 3-
cc
cc • •
UJ 0 0
uj in in
o. 3- in
o ru r~
U1 0 —
CC CO CO
o r~ r~
CC (D ID
LJ ru ru
3- 3-
t- in in
CL — —
LJ in m
o
cc ro ro
LJ ID CO
1-3-3-
Z i
1/1 1-
1 O
cn u
D —
H Q
fsj
i/i < in
>CDUJ
Z M O
O Q CC
cn
in < <
LJ 2. Q.
CC LJ LJ
O
UJ
CC CC 0
-109-
-------
> (Dr-ioj-mtDcoiDnjrucoco
o — — —-
r-
r-
Ol
CO
>
E
to
in
§
cc
LJ
0-
cc
LJ
LJ
(J
§""
E
CC
1
Z
_J
o
u.
o
Q
D
§
tn
cc
1
o
cc
LJ
Z
-
Q
tn
o
to
in
j-
ro
ru
o
cn
CO
r-
10
in
j-
ro
ru
~
8
CC
1/1
%
QEZCCU-linOX — ~> *
i o i i r- r- cn i r- ru cn r-
SOl — o o j- o —
o — — ro o r- ID
cn ooo cn r- — ~
o o o o ro r- ro i cn — m r*
to j- r- — ro CD — r- m ro o
— oruiDocno o co fu r-
cn 01 o cn — o — oir-roro
101 i r- J- o i ro to ro (—
nj cor-ru — o ro ru
o ificjito in ro cn j"
o ooo cn o — ru
o o a o m co t IJ-J-.TO
ruj-coroj-ro r-j-toiD
QDCDCor-cncn r- r- — —
ooooruiDi ij-coroo
oor-rooo m j- — co
J- O O OD — — (Dr-J-O
j-inmj-ioiD j- in m (D
i oooj-r- t j- o oj — —
j-corumin co ro ro r- cn
ojcoroocn in cn 00 cn r-
in j~ j" io in ro j- j- in to
i ooocoj-oroinj-ruj-
iniDruinromt-inrooj-
r-roco — ru oj m o cn in —
— — — ru oj oj ruru
oooocoocoj'cncocDro
ODj-j-inojj-j-r-j-ro — ro
lOrotDioininro.rcnoj-'-
rorororoj-j-j-roroj-j-in
i 1001 i iocnruJ-cn
mm ru — j- r- j-
01 ro ID cn cn ro in
r- r- cn to co co o
ooooj-ococninj-incn
tor-rotDinioiDom o
tDroroincncncooinrocnru
i— r-r-r-oarocor-r-r-cn —
Y
i Qoorururucoroj-cna
r-rurotDr-cncocoiooij-
ro — roooinoi — r-ioro
cor-r-cncnoor-r-r-coo
ro In
r- CD .r j- — ocor-rocDoru
CDfooororu — inmoror-
oooocnin i cn r~- 10 o ro
o tn ro 01 — co — j- ro ru cn
ro m ru to CD to co oo CD in o
ro j- j- ro J- j- rororoj-(D
loooij-mujfoj-i i
rocn— — — cn o r-
ooooiroojroooii i
to ro to in CD co o oj j-
ocnr-ro ru ro in — cn
ooooicncnir-cni i
ro ro ro .r iDtD rur-
j- in in j- inm in j-
< E CL in >— •— (— >- cujj-
1 I I 1 O O O < L_
1 fc fc g55sS t fe33
>->>>- _l _i .J Z Z
<<<< _J_J_JOO
3SSSoOOl55QO
Q-CLO-Q-CDCDCoooocni/*
UiLJUJLJ— I—I-JDIDDU-U-
OEZCtU-Jl/lOX--)*
O
o
ru
to
to
s!
ES
j- r-
o —
co r-
cn ID
r- a>
o
in ID
— ID
01 in
CO —
ID tD
j- ro
ro ID
m
to —
— o
ro cn
in
ss
cn ro
r- ro
o cn
— j-
m to
o r-
co cn
ro cn
j- —
CD —
o r-
ru r-
00
Sr-
ai
ru ro
OJ tD
ro r-
r- j-
— r-
cn —
ro
to co
to ro
— in
in
si
LJ <
CL >
E LJ
01 o
_l (C
_) <
5§
z <
< t-
LJ tn
N
cn T
ro o
cn f i
J- CO
— ru
o ru
ro r-
cn o
0 —
o o
ro io
cn —
r- in
ID J-
j- r-
m
— ru
cn o
oj ru
in —
ro —
ru r-
cn ru
to in
ru —
- to
j- r-
o o
r- ro
cn
0 J"
OJ O
j- in
00 —
r- cn
o ro
r- o
CO —
r- —
— ro
oo ro
ro cn
— ru
o cn
in to
3- ro
ro r-
ro
oj r-
m o
— to
m
§
=-5
-i Q
E 1C
e'l
z <
< »-
LJ tn
x
tn o o
0 0
o r- r-
> co cn
o o
!S ' S
O
in co r-
— ro co
0 0
£ ' 5
O
ro ru o
— cn cn
0 0
~ ss
o o
— i cn
— cn
o
•
ooooo — — —
oooooooo
inooo — cnj-r-ro
cnoio — ocococn
OO — — — OOO
i i oj ru ru i .r —
ooo cn r-
— _ _- O O
ro ru io tn in i r- m
cn cn o o o m r-
O O — — — O O
i i CD ru co i co in
cn o 01 CD cn
o — o o o
cn ro o o
00 — —
m — j- j-
cn 01 — —
o o — —
%.".-.-
o o — —
o i o ID j- ru in
— cn cn 01 — —
01 ™s
o o
CO I '
r- — ro
01 cn
0 0
to i —
0
—
ro cn
0 0
j- ID in
r- o
o —
r*i i ro
7
ru ro ro
CO —
cn o
o —
*
Q < E
O J I
CC O t-
LJ E
0.
§ D
O
o o_ a.
or or
* •
cn cn o —
ro co
ru ID i '
cn ro
O O
ro o ID to
cn 01 o o
0 0 — —
co ro o o
0 O — —
0 0 — —
cn m ru ro
01 ro — o
O O — —
in co i j-
o o —
r- to i ro
O Ol OJ
ru co i to
o CO -
— o —
*
i i O O
m m
CL D, DO CD
z or u _j
I I — IO
CO 00
o o
t i r- ro
CO 0
o —
i r- j- oj
to cn cn
ooo
* *
to ru if i o
— CD m o
«
io m to o
o ro cn o
i ru o j-
— CO 0
— o —
m ro o ro
o co cn cn
— 0 0 0
ro ID Ut) f*i
o cn CD 01
— ooo
— o o —
i ro 01 cn
ro ro co
ooo
fi — CD cn
— ooo
to oj o in
ru r- cn ro
— ooo
j- ' r- 01
— cn ro
— O 0
* *
o < u.
~ E h- L_
o tn t— i/i
_i _i _i
0 I I X
ro u o o
Ln o I ~
o
o
ro
r-
ru
£
o
ru
Ol
ru
ro
0
ro
o
ro
O
—
r-
o
•
i
1
ru
tn
in
tn
->
o
S
r-
ro
in
j-
ro
S
—
•
in
OJ
ro
ru
ro
ro
OJ
—
J-
i
i
j-
in
tn
in
*
o in
o —
— o
o in
o —
o r-
o —
— o
o cn
o o
— o
0 00
o —
— o
o ru
o —
— o
o r-
o —
— o
o o
o ru
- 0
o ru
o —
— o
O If)
o —
— o
O J-
o —
— o
o ru
o —
— o
O CO
O CD
o —
— o
o r-
— o
O CO
o —
— o
o o
a —
— o
z
- o
tn —
cc »-
LJ <
a. >
1/1
o
-I CC
< o
z <
LJ in
E
u_"o— oioiniomtoiD — — —
Lwcncnoioicncncnincororoco
Ljcncnoicncncncncncnoioicn
ooooooooooooo
or
o
o
— j-rococnroorocnoDoo
>incooLOr-toojiDrocnru(D
Ljoirotoinr- — — — r-roto —
Q . . . . ...
• ruinfororocom — ruj-roro
Locnj-j-ocoj-oojru — r-cn
EOjrororuruojrocoj-cnioto
cc
ccrurororooj— tooiooro
Oj-rororuroojro — j-cor-r-
Q; . . ...
LUOOOOOOOOOOOO
LJOjj-iDiOroororoincnLOr-
ocncDOicooicnoiorocD — ru
_j . . .
t/tooooooo — oo — —
gtoruJ-inmj-ojrur-r-— o
coCTiJ-inmiD — t\i r- CD j- o
cccor-furuioruiDiDaiLncoin
LJinruooor-r-roinroj-cnin
coroinj-rutDojcnro — roin
ruruoj — oj — ojiDruiDinm
t— inroiocoj-oior-.TLDtD.r
Q_coiooj-inocncocnro(— —
Ljr-inoiruor-j-cn — cntoo
cccoupr-cnruinromroioruo
Ljr-— •J'tDrurocnoiDrooiJ'
t-rum — — Lor-r-— tororu
— i
- njj-
1 i I ' O O O < U_
t/iu-U-tnoooQ—i^
Qt-t-o — Ei-Li.intn
E EOQomi-ai_i_j
ISJ
>>->->• _i_)_IZZ
tn<^'<< _j_j_ioo
>cocororo ~ — — tn in
DDDDOOOIIIOQ
ZrsiNtNJrsioOO<<
-------
r-
p-
cn
<
r
in
a
o
cc
UJ
Q-
o
5
o
in
3:
cc
o
o
z
ui
u.
o
Q
Z
O
CC
<
CL
U
CC
UJ
Q
cn
o
>
<
anzceu-iino — o
1
_
£
CO
1
ru
cn
p-
ID
ro
ro
3-
ru
3-
in
ID
o
3-
in
cn
3-
3"
in
p-
OD
ID
00
ID
o
08
CD
cn
ru
cn
ID
9!
p-
m
cn
3-
CO
cn
cn
ru
3-
M
_
00
1
r
ru
ID
CO
1
3-
ID
p-
m
3"
.
_«
3"
in
cn
M
_~
3"
3"
0
00
ID
I
3"
M
Si
1
ru
ID
p-
ro
00
3-
CD
ro
__
o
cn
cn
3-
CO
p-
ID
ID
cn
0
cn
in
ru
o
—
Ol
ro
in
cn
in
p-
00
p-
in
in
ru
in
cn
ro
r
(
cn
ID
cn
00
CO
cn
cn
cn
„
in
CO
cn
p-
io
00
00
10
in
o
in
in
cn
o
0
7
i
3-
cn
0
cn
ID
3~
o
—
p-
p-
m
cn
i
3
in
10
__
CO
H
in
r-
03
fo
(
i
i
1
i
'
O
CO
in
o
s
0
p-
ro
c
o
ID
cn
ro
ro
CO
m
in
P-
ro
p-
3
ro
cn
o
in
cn
ID
ro
ro
3
in
ro
j^
3-
3-
0
ro
00
3"
CO
o
ro
~i
ro
o
3
i
o
cn
ro
— •
0
cn
o
ru
i
0
o
m
0
(D
O
!2
O
ID
O
o
—
r- — lOJ-tocxjoioino
coconjCDCor- — onjin
tocoj-ru — — — onjoj
t
i — aiODr-oaioroo
oofoa>r*-inr-i0m
cnr-a>f\jnjQDa>nj —
r^ioiDcooor^r-r- —
QOO)OirOfUOODOr~0
tncoinmiototoinr- —
i
to f\j j" ro ro ifi i o ^" o
cn — ID ro — o) QDj-cn
r- j- ro r-
fo 01 j- o
p* OD ^ r-
oo a* o o
to r~- r- j-
to in — o
cn 3- o — o
or-3-cnr~
-
corooroo
— cn o cn —
iniDinioru
Mmruru —
— nj 01 — r- ic CD
j- o cn o
_ — „. w i i -1 i_< J" — tD CD
inocnnjfOMojror^o)
mj-j-inminj-ro —
mcn
cnnj
U3
ru
ID
cn
ID
oo
in
o cn
ru cn
r-o ro
co —
- cn
ro if)
— cn
cn
3- ID
— ru
7809
591
028
514
665
552
667
132
751
28
36
73
1586
7604
451
7650
603
6157
670
3691
416
724
451
3107
458
754
548
O IT)
ro M
co r-
ro oo
r- o
00 —
co in
oo —
r- ID
r-
f*1 O
00 ID
cn in
in ru
3- ru
in ID
oo in
in ru
ID —
in —
3- in
r~ ru
ru ID
M 0
r~ —
r- ru
in ro
oo in
3- 3-
r-
tn ru
CO 3-
r-
m cn
ru ru
ru oo
ID
o cn
r~ p-
ID 3-
3- P-
p- o
ID in
ro
o ru
en in
ru
ru —
cn o
in ID
3-
o
01
ooooooo — — —
oooooooooo
ruruin3'Cncnji3-u~icn
cnocncnooocncnaj
o — oo — — — ooo
OI:ZQ:O_JI/IO — a
t oo co cn
ooo
o cc
r- f«n
_ _ _ o o
in QOJ-J- — — o— i — —
— QTI ai cn o — — — cooj
ooo — — — — oo
oo
cn
o 3- — i oo in
ooo cn —
— — — o o
— r*- cn j- 01 o
o cn en CTI o —
— 000 — "
en p-
co ru
corooro — CTIO i ODO
CD 0) 01 cn — o — oro
o o o o — — —
j-ru — roocDcoro — fo
cnocncn — oocnonj
o — oo — — — o — o
cn cn
o o
r- — Qocococor-njinco
cnocncnooocncn —
o — oo — —
POCD — oocofoj-ifjin
ococnoooocn —
— oo —
— — o o
ooooror^j-inorofo
oicncncnooocnnjnj
o ro o — aj r-
co — o cn — o
o — — o — —
3- ID
en ru
o if) fu o
o o cn o
CD in cn j-
— — p* cn
— — ooo
iDtnruCD i ro cn — r- O)
CD — o cn — — co CD to
o — — o
— — ooo
com^-onjnjo — cno
cnoocn—• — — cnr* j"
ru
o
p-
cn
0
o
cn
o
p-
o
—
cn
cn
0
in
cn
0
o
cn
0
3
cn
o
10
cn
0
ru
-
CO
o
~
0
*
cn
o
-
!2
—
cn
o
-
ID
r~
0
in
cn
o
ID
cn
0
in
o
-
in
0
CO
ID
0
o
0
—
O
0
-
ru
0
CO
o
o
o ru
o ru
0 ID
0 0
o CO
0 0
u-Oiro — iniDODinruro
u-P-cncncncncncn3-ru
uicncncncncncncncncn
o
oooooooooo
cc
o
o
•p-3-r-j-iD — oocoin
>ruo3-inmororoin
uiro — p-rororu3-ruin
o 3 — P- o io
- co ru
o o o o
infococoru3'ior-cocn
tinrurururu — OJODOD
cc
ceo — iDcncnoo — cnr-
op-rororuru — rorocn
ccooooooo — o
cc
UIOOOOOOOOOO
Q. io 3" ro
o cn cn cn
_, . . .
in
o — ro o ID
3- in ru p- io
o o o — oo
oooo — — — — oo
cc — cncoin3- — in3-3-
oruru — incncncntO3-
"p-om3- — omcnru
cc
ujioinp-r~cn3-iD — oo
cncnoiDoo — cnro —
ro — oi — — — — r- ID
>-iDiD3-p-p-3-rucoin
CLIO — roruroruoo —
Ul — OlDODin — 03-3"
cccnp-ruiniotninojcn
uj — coiDcniDCnruo3-
t— oj ro i ru — ro o 3-
O
O
cc
Ul
0.
i o o o <
ui
in
in
UJ
cc
QI-I-Q — — — EU.J:
r: nQoaininru
V >->->- _] _l Q
CD co m co — — ^
QQOQ_I_I_IOOQ
<<<< J_J_JOO_I
CLQ.Q_CLCQCOCDOOt—
azzccu-iino — o
ON
AN! WITHOUT Q)
TANDARD DEVIA
Z
O
O UJ
• O
• Q
"£. CL
Q Q
- Z
Z <
< I-
uj in
z:
x
D
0
CC
Ul
0.
O
h—
o
<
CE
<5ZQ.ini~i-i->- ul
i i i i o o o < in
inu_u.inoot_)QiD<
ot-t-o ru.r
r EQOQininoi
v >- v >- _i _j o
< < < < _J _l Ul
CO CO CO CD — — \
DDDDOOOII-I
rslNIMMOOO<Ot—
LdUIUILJ_J_)_JDI>in
z - in
o i i i i o o o < in
— inu.u.inooooz><
i- at-i-o ru-i:
-< i: noooininru
o —
> N
i-uj >->->>- _i_io
DO U)<<<< _I_IUJ
O >COCOCOCD .-.-^
10 DDDOOOOIX-J
>-CC ZMMMMOOO<
-------
> mmocooonjnjr-
LJ ojroro — — — — iDro
r-
p-
cn
00
i
—
j_
<
Z
to
i
"~
s
o
o:
Ul
0.
..
tr
LL!
LU
0
CD
D
O
cn
z
i
o
z
z
cn
$
CC
3
0
3
z
o
cn
CE
0.
r
O
U
cr
LJ
h-
Z
—
C/1
0
CD
in
.r
ro
"
(XI
—
o
—
cn
CO
p-
10
in
3-
ro
CM
—
O
o
.—I
CE
Ul
Q.
1/1
H-
^
U")
UJ
CC
arzccu_ji/>o —
j ID ru r^- 00 co in i o
nj cn p- ro r* j- — o
j- co o m o — o ru
— ^-« M — . —
ru co m J- co cn r- i to
— ro in r- o co 10 r-
ru .T ID ro cn cn co -*•
—•
t ID o j- in ro ro i ru
— o CD in rxj j- cn
r- o j- o — o O)
— — — —
r- o y cn o — i i j-
o in — — r* ru cn
j- p* ro cn
r* cn OD ru — co i t in
in cn j- cn ro ro —
ru ru m — m to ro
i o cn o cn r* i ro cn
j- ID m cn o cu o
nj j- j- r~ co J-
i fo ro cn ro in in i o
— ID —
o ru r~ o co CD cn i cn
cn — j- co ID co j- co
rxjajj-njr-r-r^ in
i or-inmnjroco —
r- ro CD r^ o ro — o
"^
r- ru r- ro — ID 00 J- cn
ru ID o cn j- r- ro ru
ro ro in nj p- r*1 i"* Lfi
iD P- — cn tD o i iDifi
in CD o — cn j- njm
ro ro ID ro ID r ro
i ro co r- i in — to ro
ro in .y co co ru
— o cn ID i in cn i ID
cn ID m — ID GO in
ro ro in ro r* ID nj
r> ru cn cn i — m i o
ID ru ru ro 10 ru ~
ru ru to nj ID to co
<3-Q_t/)(_(_(_v
l I t l O O O <
U") LL. U_ l/l O CJ O O ^
o »- »- Q — — — :c u.
£ i: a o a i/) i/i
OD OD 00 OD "•• — '
NNISIMOOO<<
QQQO_I_I_IOU
0.0.0.0.00000000
Q3CZCCO_lcnO —
ro
nj
in
ru cn
(XI CO
cn (xi
0 CO
3- ID
CO 0
— ro
CO (XJ
— ro
— in
CO (XI
cn p-
ro —
cn p-
^j ai
U (XI
co ro
cn ru
(Xl — "
in ro
in 3-
(XI —
— CD
CO ro
CO (XI
in cn
— ro
in aj
in j-
— ID
CD —
3- X
ro co
in ~*
CO 00
- (XI
in CD
m ai
r~ CD
p- co
3- —
O —
Sffi
z
o
•— •
- <
o —
>
t- Ul
3 Q
O
I O
>- a.
«_. 2T
z <
< 1-
ui cn
r
fsl
CO P-
— in
cn ro
CO CD
(XI (XI
in ro
3- 00
3- o
cn nj
ro —
cn p-
(XI
r~ CD
r- r~
ro —
in —
oo cn
3" (XI
cn r~
ro in
(XI — ~
ro o
(XI 00
(XI —
rT)
CJl
P-
0 p-
r- 10
3- (XI
— ro
CO CO
in ~~
co in
CO O
3" (XI
in o
in co
3- —
p- co
3- ro
ro (XI
3- (Xi
3- (XI
o —
CO CO
3 (XI
z
0
5
~~
•"* ^
— Ul
• Q
• Q
r cc
• <
Q Q
- Z
Z <
< 1-
ui cn
T
•x.
cn
o
CD
in
3-
ro
—
(XI
™
o
—
cn
00
p-
CD
in
3-
ro
ai
Cl
o
—
CC
LJ
CL
Nl
o
t—
0
h-
CC
— ryro — — — — oro
ooooooooo
oo — o — — — oo
03CZCCO_ll/10 —
j- cn — in — ru — ro
0 O — 0 — — — —
* * * *
cn cn j- — ro j- r*- i j-
njinfuinruro— to
oonjo — — — o
i co cn ro in ro T i co
r- o in — nj — o
o — o — — — —
* * *
o j- r- nj ru — i i j-
in r- — in j m —
0 0 — O — — —
in m CD co j~ — i i cn
ID r*- ro j- ro ID r-
o o — o — — o
* * •
i in cn m — ro i j- r-
j- co oo in in o r-
0 0 0 — — 00
•* *
i co ai ro ru cn ro i y
ro ru r^ ru ro y CD
o — o •"* •"* ~" o
* *
m o — — ID ro nj i ro
in — cntDinr^j" —
o o o o — — — —
* * *
1 1 O 0 1 1 1—0
ro GO o cn
— 0 00
* * *
ID — r* j- cn ro in i .±
mj-minrinj- —
0 O 0 0 — — — —
* *
i (DJ-roio — cnroru
r** o ID ru ro *™ o co
0 — 0 — — — 00
* * • •
— coinmcnmco — cn
(DtDcnmroj-roocn
1
* *
Pprocnc\j- i gru
00 — 0 — — 00
* • •
i —3-00 i r^ CD 3- (xi
o o o — — o o
• *
(xi CD P- CD t — in i 3-
CD (^ — CD CO 3" in
o o — o — — o
• * *
co co ro ro i p- cn i o
in 3- p- in 3- ro co
o o o o — — —
-l->-
l l l t O O O <
LO U. U. LO O O O O ^5
ai-t-o ' — E u.
z r o a a c/> cn
MNIIMrslOCO<<
oaoo_i_i_iuo
0.0.0.0.00000000
QtZCCO-lc/lO —
o —
0 3-
— o
0 —
o ro
— o
o ro
0 CO
— o
o m
o nj
— o
o —
o 3-
— o
o ro
o 3-
— o
o ro
0 3-
— o
o —
0 f
- 0
O CD
o m
— o
o p-
O (M
— o
O CD
0 f
— o
0 P-
O (XI
— 0
o r-
o ro
*
— -• o
o ro
— o
o 03
— o
o —
O 3"
— o
O 3-
o in
— o
z
o
h-
O —
>
1- LJ
D Q
O
I O
i- cn(McncD3-in(Xi —
U)Cn3-ODX(XI(M— t
•3"oco(xiroro — op-
cnp-corxip-cDinp- aa
n nj —
CE
SLn(Xioroo3-ro o
ro — co cn co r~ o 3-
CC — — (XIOOO— (XI
LJOOOOOOOOO
Ljcnnj(xirop-p-ro co
o (Xi o 3~ LO o o cn —
cno — — o — — oo —
CECOcn3-rop-inx 3-
oco — in(^ — 3-co in
CEincoco3-3-3-in CM
uicDcnmcD3'rorvjoiri
r- OD r- in in j- ID 3-
t— (xi o — oor^coo r-
CuCorojp-r-cDfxi oo
UI3-— (xip-ocnm ru
cc(MtMocncocotnoro
uicorur- inoo —
K- — (XI — — (XI OJ —
Z i l 1
_
I I I 1 O O O <
cnu.u.cn CD X CD OD — —
3ODOOOOII
ZNMNMOOO<<
OOQQQ_I_I_IOO
in
uia.Q.Q.U.Q3CQLXloO
oruiuiui'jj_i_i_iDD
0
Ul
-112-
-------
CCa.>-oooooooocn-"0i i r- KI IP —
> o o o o co fu — to cn ID KI KI ip o ru
i a> r- cn ru j- — co cp o m o ru in r- —
S;riPKiQDr-r-o5KiruKi r- KI 01 —
iKiruAjrCruruKiruKiKiK) ruajKij-
— (U
isii
nj OOOOOOOOOI0CD
o o o o o i a i i
ru m S ru aj KI
ooooooooo
CD K)
in —
Kl Kl
j- ru r- i
— _ to |
r — r- f
§ — OOOOOOO — J- O — y r~ in ru IP
i (0 o u5 r (^
i ru ru nj
IP ro —
i nj ru ru
ooooooooocr>CDCp~"t0
ruoj-oo — j-ruruKi — ipKij
co — inj'r-10-- — — oiioipr-r^
Kl 01
nj —
O) fu
O Kl
— r-
KI r-
S"1
ss
K) (*•
--ruru—• — — — — — rururu — — t
!5°aoppr;iP i r; — cp o 01
auioiiniDcor~ininTJ'iniDJ1 — ru o —
G — — C3OOOOC3OOO — — — — —
GOOOOOQOOOOOOOOOO
goi — r-r- — r^r-iPooi — oiKir-Kimr- o to
coruococncDcncncno — ococoai — ru o —
o — — oooooo — — — ooo — — — o
cco,»- o —
o — ooooooo — — oo — — — o
ru j— KiryKiruKiip — j-j-i ir*r-—KI oru
— CD — —fliocooioioi — — moo— o —
o— — o — oooo — — Q_^_. _0
* • * • *
— Kicomru— lOKiruiriKiicDjpKiKioi ouj
o — — o o — oo — — ooo — — — o
o moepTCOicoKi — conjKicpKir-inr- 010 u.i
~ r~i0Kicor- cDoioio — — r-ipcnruo oru u. i
a _ — o o ooo — — — ooo — — — o
• • • *
o — — ooooo — — — ooo — — --o
CO KICOOI i i iru(0i i iruoj-roru OKI
CD o o cnco —COOOIM o —
o — — oo — o — — — — o
r*- —KiipruKij'jgujruoicncDtDfUin — r- oj-
o— ooooooo — — — ooo— — — o
10 r--jgiintrur-cno — mr-coiPtoio o-
00 — O CD Ol oflDCD — — OOTQCnoru O —
o — — o o — oo — — — ooo—•— — o
in cooiru — Kir^irioooir^r-KiKioKij- OCD
o — — oooooo — — — oo — —•— — o
o ....
oooooaoooooo
toot
(S»!
ii
> O O I
> ffi cn
f — fu
ru KI nj KI
— — in T
KI ru KI nj
r* — r- (D
ru aj fu ru
>oooooGOioii0cfior-i i —m KIO
>(Qoocoo>cnin f f o \c f o t cnj-
> KI IP o — r- — o CD — in KI co iPoi ODIP
i in — — 01 — — o njKtr-oico — rv — KI
r- r- i - r^
p* r* in K)
8 i o f^ i
— !/)>>>(
i^yit/i
ip:j:j
>- z(to:x>->>- _)_)_j;
Su^Z^JCOCDcScD — — — '
oftSi-Ecoooj-j-joou;
222;
in - In -
II
UJ UI
SO1U3J-— inj'flDj'or- i r^ CD co ID CD
— — oior-ocnco — o CD CD CD o M
o — — o — o — oo — — ooo — —
i i—Kir-iru — oincoi i
ocncn rurur^cncD
— oo — — ooo
o — — — oo — oo — — ooo
* • • •
lf)U3OOlllD.ylPOiCDl0iO— i i
o — oo o — — o — — — o
i/i_j_i-j - njj-
i O O o i i i i o o o < u.
1/1>>1>Q.I/IU.U.I/1LJU(_)O_JD
rixi>-z rooQt/i»-i/i_/-j
>- ZCCQ:>->->->- _i_t_izz
<1/1OUJUI<<<< _J_i_jOO
Df^ccij3D55oooiiiQQ
QttCD<-EQQQO_j_l_ic_)OOII
<<<_J*-'<<<1l/lUl
Ctoooooooooooo — o o o o
uiooooooooaoooooooo
uotor-CDCD — 3-j-ocnincoifiaDoK>cD
Cur-— ipoij-cJiruii — — ru m ru o ID j- r-
OCno)tDP~C7)£JlOlSDlOOtfl — OCOfuru
t/iooooooooo — —o — oo — —
o —
— o
-r-— r-— CDUJorutfiCDruo'ico
- — ocoj-in — 01—ipcooojo
icoiDr-ioiPOiruinnjKi — IPK)
jKitoinj-ruiflcnKiooo — KI
•— rur-cocnior^ipr^inr-oo'
• ? "f1
i tfi °
ru — — —
• cn KI a- o cn i
• s- ru j- — 03 i
o o
. - ru r
— o
1/1 —
CC t-
UJ <
- i
U l/l
Z CC CC X- X >- >-
:QCCZljrD§§OO<
JOD
-------
- S
2686 23468 28599 27707 86868 R
5820 23200 - 25260 27730 0
4400 28000 32713 29437 28646 8
7081 28650 32438 29695 28998 U
3300 21600 25400 24800 25000 F
3008 23736 - 26853 - 0
4085 22758 27894 25818 26061 M
3575 23292 - 25836 - N
4049 22086 - 27702 - R
6713 25425 28596 30375 27057 C
6404 25851 29553 29994 26907 L
K!|
cn o r*
01 in CD
cn — r>
3 21758 33975
eil 70 23230
22933 26935
— i" j-
ai 01 M
O — ID
ru ru ru
cn o CD
M ro in
— ro ru
r- co ru
CD O CD
t-~ o —
J- CD CD
a> in j-
r- fu ID
o co ro
ro o —
&c|
?§s
fi ID at
k DZUBAY MDS-S 1
k RODES DICOT 1
i BURTON CHAMP 1
U& Si
111
£ o -
So ro
o ru
Issss
to in ro in r*-
ru CD o OD m
— o o ru ru
cn o r- i ru j- i i
cn o j- — ro
CD r~- ro jo
lO ro CD o o
ru — — ru ru
toco — 01 ru co in in
r-rurunjru — j-j-
ojrurufururururu
m o
-. ID
ru ru
ru o j-
ru o j-
ru r> CD
ru co ID
if) o rj)
— o ru
ru o —
CO — ru
ru o —
0 0 J-
ro m co
f O ffl
¥ ? to
\ BURTON CHAMP 1
r MULLER RSP 1
\ DZUB*Y MDS-A 1
UJ U UJ
o ru — co —
ro o o j- j-
ru ru ru ru ru
r* r- cn 10 o
in co ru — cn
o m fu r- ru
OD co r^ 01 cn
co ru 01 01 in
3- in — 10 ro
in o — ro CD
ru — o ru —
a) ru ru nj
ID ro — 01
— o — ru
ID O O 1 f
— — ru in
r~ in — in
o cn 01 o
ru in cn o ru
!!£5S
t± t± Q H y
EGO
< < <
S S 5 o o
&&&S3
£
ru
s
j-
o
ro
r-
S
ru
S
p
p
CD
24357
r~
ru
cn
'
£
ru
cn
r*
P
0
o
i cn o
01 01
3- ro
i m o
ru co
— r-
ru in
ru
i ru o
«n
r* j
ru
1 0 0
in r-
co ru
O CO
ru
ru m
1
?
o
0
21000 28800
ru
o
r-
to
o
O
j-
ru
s
o
o
p
o
ro
' UCD CAM ILL SMDAY ]
s?
si
ro o
j- o
r ro
j- o
j- to
ru
in o
r> m
CD r>
— ro
ru
co in
ru
II
ru
ro o
?5
o ru
r-. —
o ro
£S
CO ro
r- o
o o
in
tn ru
_l O
_i u
ru
R
ru
ID
te
r-
in 01
O ID
ru ru
to ro
ru
co ro
00 ID
ro m
r- ru
ru
s?
o ru
ID CO
r- ro
?£
ru
o r-
ru co
cn in
j- —
j- j-
cn r-
o ro
CD CO
in —
cn cn
o ru
co r-
ro CD
ru r
ru
in o
££
ru —
ru
01 CD
ru r>
f —
ro ru
ru
10 ro
GO ru
in ru
CD in
O J
— ru
r o
CO (D
ru o
o ru
MEAN (ALL SAMPLERS) 14040
STANDARD DEVIATION 1577
to —
to cn
ro o
S'
8$
to ru
ru
in oo
J- CO
ID in
co ru
ru
o ID
00 ID
ro ru
ru
ru J-
££
in -
ss
— cn
in o
0 CO
in
j- to
ID in
01 in
r- —
M r*
ru cn
ru in
ru
(D ro
CD O
01 ru
ru ru
ru
ru o
01 CD
ru "~
m m
co ru
cn o
in j-
ID cn
o ru
oo —
r- r-
o —
ru —
r- r-
01 o
cn ru
MEAN(OUFOMLGI ) 13715
STANDARD DEVIATION 1653
>
o
s
o
>
<
ID
in
3-
ro
ru
0
01
CO
r-
ID
m
j-
ru
8
or
rsj
0
o
cocorur-foocDiDj-j-j-incOJ-tD
cor-j-co — cnioinj-m.rinin.rfo
o — — o— ooooooo— • — —
ooooooooooooooo
in— r- CD 01 — — 3-rocococoroj-co
oo — — oo — oo — — — ooo
ro 10 cn — in i 01 i i ro ro j- i r- ID
OOO — Ol 01 OOO ID ro
__ _ _ „ o o ___ oo
omoocncncncno--"" ' co ru
— o — — oooo — — — — oo
CD i ru ru r- i ID i i CD ru cn i to 3-
(j) __ _ [Q cn Ol O Ol 01 —
o — — o o o — o oo
• • • * *
oo — — ooooo — — oo
uicor-moir-j- — j-njo— i 010
coooi — cocooicnoi — — — oro
o — o — ooooo — — — — o
• • * * *
coco — — oooiooicn — — — r^CTiro
oo — — oo — oo — — — ooo
* * • *
j-cororu — coint~~fO3-iDcDru — CD
r-comrocococncocnooocnoio
inr-cor-ocororuinru — coj-roro
OToio — r- 01 o 01 o) •— — o 01 o ru
oo — — oo — oo — — — . o — o
* * * • •*•*
ro j o o — ro i — o t i icnoij
o o ru nj ID r> 01 cn ru o —
— „ __ „ o o oo — — o
mruoocnioor^inr-r-to~:rto
croiorucnoiooioiooocnoi —
oo — — oo — oo — — — ooo
CD Ol — — O OCOCQOOOOOl —
oo — — — — oo — — — — oo
rucooicnooJr^oirucoj'.yofuro
oico — — cncooicooiooooiruru
oo — — ooooo — — — o — o
* * * * •
j-roinTCOoroo — ru r- i rojio
o j ru ro Ol CD •— o 01 — o ocrifU
— o — — oo — — o — — — oo
cnr-oj — 0100010 — — r-oiro
oo — — o — — o— _«ooo
inj-o — j-cnj-fooi i ro CD ro cn in
oi — j-rocnr-ocnco oor-r-fo
0 — — — OO — OO — — OOO
• • • *
QQODOOOCPOOGl — — — Ol CD .T
cn »— o_ OL <£CLUII-I-I— >• tn
101:3; i i i i o o o < tn
i/iu< o ce x: i: Q Q o in cn ru
>- z z or >->->> _i_in
cnroiDtuocDru3-ocn — coruj-
ujcor-ocDOicnoior-r-rumrjico
•cn — ic~-r«-ij-ruror-iDr«"ic3ior-o
cncocoorjijiCj'r-cor~-r~-rucocn
Erj)omoruiDr*cor~Q3CDcninr~
or ru — — ru — rxiru
CCOOOOOOOOOOOO — 0
Q; ..... . .
LJOOOOOOOOOOOOOOO
UJJ-M — cotommin — J-coruj-o
ctruinMco — mruruinonjocDj-
or"'ooocnoio)oic'ooo — co
ui — — — —ooooo — — — — oo
cccn~ruor-r*ruinrurur-o — m
o — rocDj-ocncnr-moicoro — j-
r-cofuj-rua-oj-cor-inirino
— — ' * fu —
o . .,.*..*.* T. . . .
or— — ocncoco — cnr-cocQiororoo
ujroior~oocor-r-r-J-coooo
»- — — cororocoru — rurur- — — ID
>OC£ I I I i O O O < 1/1
r-j
>- ZZCC>V>->- _J_IQ
cnoou<<<< _J_ILJ
5occordDDD3oooiili
§rNJOZ3DDNJf^rs4rsjOOO<
-------
r-
01
CO
i
>-
x:
ID
i
in
rjr
UJ
Q.
o:
UJ
r
o
§
(J
O)
3:
CL
2
Z
5
IT)
afi
in
UJ
o
CE
_/
0
Q
s
t—
U")
z
o
in
0.
K
o
O
or
LJ
Z
D
0
*
ID
in
r
ro
ru
_
—
o
cn
03
r~
10
in
*
ro
nj
*~*
D
O
tY
LJ
0.
in
ESULT
oroonzrro-iino —
ooruooiD — 3- 3- (Ti i r-
10 ro r- oo — — roj-j- ru
ooruioruioru.ro ID
— — _ruro — rororo ro
ODoiOJ-inruj-r- — i 03
§r- ro — toruruiDo ru
o 10 ro cn — • r— cn to f
— — 3- — ru ru ru —
oo i oooruincncn i ID
cn ru T o in to ID ru r-
j-ro — o j- — ro — 01
ru ru ro — ro M ro ru
cno — orur-oroi iru
oooruinj-in — ID OD
— — ru ru ru ru
oo — r^ooiDroj- i i in
r- — f— cn ro r- on — 3-
in fi r^ OD to in in cn cn
— — — —
mo i o r- o r- — i cn r-
r- .r ruomcnnj loru
— — — ru ai —
r- o t cncn - cn in in i o
103- ro 00 in oo J- r- r-
IDOO rooiooruru in
loor^ooroco — oro i —
oo — r-oitooiruoo ID
in ru .r l ' r ~ ^2 2 0>
r^ o i i j- ro i i i j- ru
o ID j- — ro
— — ro ru ru
— ooio — o .r J° r- i r~
r-r~t-ro.rj-oio.j- ID
inioooiorooorororu r-
— — ru ru ru —
r- o t o — (nmtocn^-ro
rucn — — in ru o cn in o
— — — ru fu ru —
ruo — ID — cnrooo.rrur-
oor-ODODOir-oiru — — OD
lOocnoinODruroru ip
~
cnooD — ror-ooo i com
rumiooomooru r-io
loroo — aocoooi o
— — — — ru ru —
ino i cn .r — t in ro cn 01
ooin — in ru into 10
ru ru
ro o co o r- ao i inr- i 03
in o r- 03 r- j cnto 10
j-ru — oiooi ruo r-
— — — ru ru
01 o — to r- r- i roin i o
roinoiDoo— oor- ro
in 3- co ID 01 r- cnoo 3-
01
101 I I I O O O <
inoinu.u.cnoooQD
r o ic it Q D D cn cn
>- >->->->- — i _j
•^ CO ^ ^ ^ «*" i _j
QO IJLJ Q) CD OD QD ~" *— •
3O3OODOOOII
ooraaaa_i_i_/oo
Q.Q.O.O.O.Q.030D0300
in
in
j-
ro
in
ro in
ru in
j- o
ru -
ru cn
r- ro
cn ro
0 00
in in
r en
ru
ru in
o r-
r- r-
—
in in
ro 01
— r
fu in
cn —
3" ID
in —
r- to
cn t
ru ru
00 ru
CO 3"
cn cn
— cn
ru
— r-
ID O
3 r-
o ro
r- cn
in 11*1
u ' lu
r- f
in —
J" (0
in 03
—
ID ro
3 03
in r-
OD 0
— r-
o ro
ID ro
"™*
Z
o
— <
o —
t— UJ
D Q
O
X a
t- CL
3 Q
- Z
Z <
< l-
LJ in
ru r-
— 0
f ru
ru —
ro ro
? 8
j nj
ro cn
O ID
r- in
ru
cn o
r- r-
—
r- ro
ID ID
— r
~
ru ro
in —
— ^
cn o
cn oo
Cn IO
in ru
r- —
in in
— i*j
cn in
cn in
"
— j-
U3 0
ru 3-
o o
in r*
o in
—
— in
a?
—
if) M
nj CD
ro r-
O CO
— r*
to nj
in OD
ru OD
"" "
s
<
~
— UJ
— Q
z or
- z
z <
< f-
LJ in
o
o
in
o
*
10
in
j-
ro
ru
„_
—
0
01
CO
r-
(D
in
j-
ro
ru
-~
O
0
or
LJ
0.
M
O
O
<
cn r- r-
ru in ru
j- j- oo
— 3- —
000
OD r- 3
3- r~ ID
o o o
or o o
j- ro ro
3- 3 in
o o o
in j- ru
j- in ro
000
o in t
ru Oi
0 0
ID - 3
ro — in
o — o
•
o m OD
in — S
o — o
in r- i
(D en
0 O
•
oo cn i
ID 00
j- cn -
r- ru ID
0 0 O
*
CO ID i
f. .
0 0
cn j- o
ro — to
0 — 0
• •
3- 3 1
ro 3-
O 0
r- ro r-
t r- ID
* •
ID O1 OD
3- Ol t-
o o o
ru r~ i
10 ro
o o
oo r- cn
ro — cn
0 O 0
*
10 cn cn
j- ro to
o o o
*
in i- <
I O I
in o in
T. Q T.
V >.
< in <
00 U OD
D Q 3
MOM
Q or Q
Q- 0. 0-
UJ UJ UJ
r~ — o —
ro ro ru —
ID 01 f t-
ru ro
0 O O 0
cn r- cn ID
to ru ID 3-
o — o —
r z o: o
— ro r- 3-
— ro ID ro
— — o —
• * •
r- ru r- oo
ID in in ro
o ru o —
oo ru cn cn
oo ru in ru
o — o —
01 ID ID ro
r- ru in in
0 — O —
• •
01 J — 0
r- y in 3-
o — o —
03 3-
j 01 cn uj
o o o —
* •
in ru r- ru
ro — 10 —
0 - 0 -
— ru cn in
— o ID r-
o — o —
* •
i ro o i
to o
~ 7
J* Oj CD CD
j- cn m m
0 O 0 —
•
O (\J J- CD
01 <\j r- j-
o — o —
* • *
in j- o ro
r* o co in
* •
— OJ o ro
CD ro r- m
o — o —
ro, o — i
CD •" cn
o — o
* * *
— OJ O I
O> J- CD
0 — 0
* * *
r- tn ru i
in CD co
000
*
r D- cn K-
l i l O
u. u. cn o
t— 1 — O *— •
t Q
>->->-
•*£ ^ ^
CD CD CD
ID 3 13 O
Ni N NJ O
Q Q O _J
CL 0- CL CD
LJ LJ Ld _J
,\i ro O)
m
cn o nj
— ru o
o o o
nj cn r-o
to _r o
— — 0
_i en o
ru to i
j- 01
— —
- fO i
in ^
CO CD i
ro, f\j
— —
("U i i
—
01 i '
10
—
ru i in
UD 0
— o
OD — i
ru fo
j- cn i
01 in
— —
i i —
°.
0
ru r i
10 in
— —
ro o ro
in j- o
— — o
o ru —
ID in o
-~
ro i 10
to o
— o
ro in in
— .- o
j- in i
01 r-
— ru i
— .-
o o <
0 O D
Q o in
_J
j
§O I
o <
_l _/ 0
_l _l Q
03 CO 0
_J _/ D
01
ro
0
J-
o
3-
O
-
-
O
in
—
ru
r-
o
OJ
ru
—
ru
ru
—
•
ro
00
o
ru
03
0
•
cn
in
n3
„
*
ro
*
_
Ol
—
*
to
cn
o
•
cn
0
r
00
0
cn
o
in
10
0
cn
o
ru
.,,
Li-
en
_j
I
i
o
Q
O
D
o cn
O 3
— 0
O 3-
0 f
— 0
o oo
O iD
— o
o cn
o ro
— o
0 ID
o r
— o
o j-
O 3-
— O
o —
o j-
— 0
o r-
o j-
o ru
O 10
— o
o ro
o ^
— 0
0 00
o j-
— o
o j-
O T
— O
o ru
o j-
"
o o
o j-
— o
0 1-
— 0
0 0
O ID
— o
o ro
o to
— o
z
o
— <
o —
f— LJ
li
" Z
z <
LJ in
r
u.rooiOinoin — roru o
u-cnooru — o — iDinro cn
ujj-j-incnooCDCncncn r-
oooooooooooo
0
o
• ~~ r— 01 01 oo r— to r— 01 r—
>cnr-ior-roooroj- —
bjcnoiino3r-03iDr~— r
inoD^rur-coincncnj- in
JC — inruruiooi — — ru in
Binr^r-ooo — ror-co o
cncninminro — oro 01
ororu — — ro — — — — ru
LJOOOOOOOOOOO
ujrur-j-oiofoooocn r-
O-oor-rooocnroruiD cn
o 01 ID 01 01 r~* 10 ru 01 o ro
inooo — — o — — — o —
or o ID oo 3" CD OD QO ro ID —
oj-roin — onjrooioi oo
ccocnr-ODOj-oicno —
LjroiDCDinto — j"r~roo
inr-roror- — coioru to
— j-oioiinru — — ru 3-
}— j-ocntooooioroo j-
CLoocnroiDtDj-iDro —
ujcnrooj-inr-r-cno ro
ccoj-ror-inror-otooro
ucorocnru— — in oo ru
i-roruj-r-r- romin in
Z 1 I 1
i o i i" i i o o o <
O"-Q^»— O"-"- — 3CU,
>- >->->->- _J _l
in ^ irt rf ^ rf rf j ,
> CD UJ CD CD CD CD „ —
3QODD3000XI
ZMOMNMMOOO<<
OOOrQQOQ_l_J_JOO
in
LdCLCLCLCLCLCLmCDCDOO
QTLJUJUJUUJLJ_l_j_IDD
rroorzcco-iino —
-115-
-------
— roroai — j-j-ruio
____f>j — ruru —
INTERCOMPARISON STUDY OF TOTAL SI . NANOGRAMS/CUBIC METER; PERIODS 1-16. MAY 11-18 1977.
SULTS PERIOD 1 8 3 •« 56 7 B 9 10 1 1 12 1 3 14 15 16 AVG SO
o CR cu oj CD cn ru — t*i o \n
r- in co ID o — ru cn co M ro
o i o o en J" ru i ro ID J J-
ru rur^uor-ru o — j- ro
ro rutDoj-— nj j- — o")
ro CO ru o .T o roocoin
CD — r- co o — rofAJj-r
r- cn io o ru ro cnoruro
rororuinor- iDf-CDin
mi m in r- co ru rocor-ro
oooocomi ij-r-njM
^ ru — r- j- — r-iom —
ru uo r- T — ru r-rji — ro
to in r- tn — — cn m cn o
in r~ HJ cn in ru .r ru CD r-
j-ror-njro inco ru iD
iPCQtOCQO — 3-C- J Q CD
ooooiDcnoru— cooj"
nj — rururoj-j-njj-rooitD
i iooi i i co — ID m in
or- CTI r o ro r-
nj ro CD ro j- — ID
cn ID 01 o cn ro ru
O O O O ID J" — in ro 01 O O
r- cocn — cnroj-iDcncoinr-
cnrucotDiDcncnj-coincccD
i aoomcojioiD — 100
j- — CD CO f~- — ro — — y CD
mo — io ru ID .y — 3- ~- in
co cn ID — fu nj ID cn ID ro ru
co j- cniDcnror-o— mo
j- 0 OCDCOJ-OJ-— mru
03 — — ruru
OOOOID— i oocoroo
uo co ro m — ru ro — ID o ro
roor-iDcnm m co cn r- cn
cn cn J- ro m r- ru in
in CO cn oj-roro —
r~ co ID ruioinruj-
ooooiruot— t~-fui i
uO r- ro ro j- r- ro — m
3- Cfl 3~ r* — ro — CD ru
upcooin ru m to CD io
OOOOIffl— lOUll 1
o r~ r- cn r- r- r- j-
m .r ro in — 3- oro
i l l l O O O < U.
>->->->- _l _l _J Z Z
m ai m co dd^S°
CLQ-CLCLCOCQCOOOOUlUl
LJUJLJUJ_l-J_iDDDLLU.
O
ro
CD
cn
to r-
cn ru
r- co
in cn
in —
r- r-
3- —
in KI
o ru
in ID
r- ro
KI r
co ru
o o
S?
cn ru
m CO
o 3-
o —
y ru
ID CO
r r-
— ru
o ru
ro o
m —
ru in
CO —
CD r-
cn ru
ID J
o o
— ro
CD 0
in o
cn ro
ru —
ru —
cn j-
o nj
cn in
in r
CD ru
J- a
r- in
ru —
z
— o
Ul —
or t-
UJ <
< Q
Ul
MEANIAL
STANDS
f\ C\j
01 ro
ID 0
in ro
01 in
o ru
— o
0 3-
CO ru
0 (U
££
O CO
CO ID
to cu
— m
in r-
ro —
ID ID
ID r-
o —
co —
o cn
cn —
-r rn
nj o
cn ro
— m
o r^
00 -
— in
ru m
ro ro
in in
cn M
0 0
s;
r- ro
nj CD
cn cn
cn in
r- ru
ro CD
r- o
ru CD
ID ro
Z
o
MEANID.M.L.G. I ,K
STANDARD DEVI AT
a
0
in
ru
CT
CO
r-
ID
in
ru
Q
O
or
Q.
o
o
o — ru o
o o o o
o o o o
Q r z ct
O O O O
o i — r-
r- co ID
o o o
r- CD io
o o o
ID r- co in
o o o o
if) cn ru in
o o o o
* •
I CT) O CD
o o o
r- — r-
o — o
— ID in co
io 3- m in
o o o o
i i — fO
cn LD
o o
in in o ru
io r- CD ID
o o o o
i r- ru ro
CO CT) ID
O O O
in j- i j-
ID co in
O 0 O
r- in — o
ID CT) CTI ID
o o o o
i ru ID in
co cn r-
o o o
in J- o r-
r- o — u3
o — — o
j- r* cn ro
r- in 3- r-
o o o o
Ul U. U- Ul
O i- t- O
2: n
rsj N IN f j
O Q O 3
0. (L 0. Q.
0 0
ro r~
u _i
ro in
o o
o —
— ru
o r-
o ro
ro m
CTI O
O —
o —
01 m
— ru
ID ID
o o
ru ro
ro
i ro
ID
O O
U O
Q a
8§
_J -J
-J _i
CD CO
O O O O O O
o cn m CD co r-
— o — o
Ul O I —
— 00
cn i OD r-
o cn ro
— o o
ru co cn
— o o
- ID
— O
— o
i r- — CO
o o o
j- ID o tn
— 0 — 0
CD J ID —
— r- o co
— 0 — 0
i co ru fi
cn o cn
0 — O
* • *
cn — m ru
ru ID o ID
— o — o
cn to M in
ru iD 01 iD
— o o o
— 0 — 0
i r- ru ro
co cn ID
o o o
co co j- in
r- in ro j-
— o — o
U3 — ro ro
tn r^ o i-
— 0 — 0
m i co o
cn ro in
— — o
O < U-
CJ O _t ID
— H l- U.
_l _1 _J
_J _J _J
0 I I I
O < < <
_J 0 U U
_i a a a
-> x
3- O
ro j-
o —
CD r-
co o
J- UD
ai ro
M CO
ro o
— o
ro r-
nj —
cn in
ru fu
m —
CD ID
3- CO
ro nj
ro ID
ro m
ru 3-
Ul Ul
o o
Ul Ul
§§
I X
Ul Ul
u. u.
o r*
— o
0 CO
o -U
— o
o —
o j-
— o
o ru
o ru
— o
o ro
— o
o r-
o nj
— o
0 CD
o j-
— o
— a
o j-
o in
— O
o ru
o ru
— 0
o —
O 3-
— O
0 CO
o ru
— CD
o —
o j-
— 0
o ru
o ro
— o
o m
o j-
— o
o ru
o ro
— o
o r-
o in
„ o
z
£5
< O
Ul
i 1
z <
< 1—
UJ Ul
r
§
o
g
o
i
Ul
£
or
cc
g
cc
LJ
£
O
Ul
o
cr
cc
a,
£
u
z
rxi
>
Ul
u
o
coj-rororo — j-cnorooj-
oooooooooooo
r-r-coinior-cD— ocoroo
ror~~ njroror-CTicn- cnin
— rocoocoj-cDfuinoDrucn
— L£> — CDCDrOCnoCOfXJCTl —
inror-tDfuruoinrucom —
tOiDCDococomruro — coin
j-iDintD — ocoo — corocn
oooooooooooo
rurur-in — aicoinr- cocnro
ID — J-cocor-cnj- — r-com
iDO^coincncocoincnrocDCO
oooooooooo — —
(DCnru — rorucoo3cor^--in
cnocn — r>ruoiiniDtDCDCT)
jiDiniDru — cncn — uDincD
inr-rucnrum — oio — r- —
cocouDcD
CLQ-Q-CLCDCOfaoOOUlOl
a s: z or o .
-116-
-------
in r- — r- -cnruMro
rurutxjruru — rururo
p-
p-
cn
cc
LJ
u
co
o
I
o
-
i i t i o o o <
(/1L.U.U1UUUO3
Et-k-Q — — — ZU.
z a a a cn cn
OODOOOOXI
NlNNjrsJOOO<<
QDOQ_I_I_IOCJ
<<<<_i_i_iao
Q.CLCLCLCDCDCDUO
LJULJUJ_l_l_J^O
z cc -
LJ <
^>
3°
Q
_l CC
_J <
< Q
Z <
< t-
m rn
P- 3-
co p-
ro -
3- ID
CO ru
ru —
ru ID
ro 3-
ID r-
3- 0
3- ro
— ru
in in
ro 01
O CD
§5
o ID
ro co
— CO
ru co
LO in
p- CD
3" O
CO ru
co o
o in
10 cn
— ro
ro ro
3- ru
cn ro
ru ro
o r-
3- ru
ru ro
o LJ
• Q
—i
• o
^ cc
a n
z <
< t-
LJ in
Q
l/l
ID
LO
Q
o
CJ
CL
o
t—
o
ru
o
j.
01
o
o
s
-
in
O
*
i
ro
o
*
r-
cn
o
*
$
O
*
K
o
*
in
CO
0
CO
10
0
CD
p-
0
'
ru
o
o
CO
o
5
^
*
OJ
O
_^
*
£
0
^
i
1/1
Q
n
CD
D
N
a
0.
LJ
a)
o
j.
o
-
r
CO
cn
o
*
cn
in
-
r-
co
o
cn
cn
o
•
CO
o
o
ro
_
r-
r-
o
cn
in
o
'
s
-
cn
CO
0
CD
*
_ .
cn
o
*
o
ru
^
3,
O
*
ro
_
*
^
I
1 1
K-
§
N
Q
0-
LJ
ru
o
[-.
ru
-
z
CO
3"
-
O
in
-
CO
3-
-
m
3"
-
ru
o
*
o
ru
_
CO
o
_
*
8
o
*
ro
r-
o
0
r-
-
s
-
cn
-
r-
—
_^
ru
^
•
O
J
*
co
in
_i
o_
i
1 1
H-
m
U
Nl
Q
a.
CJ
—
o
in
CO
o
cc
cn
cn
o
*
ro
in
o
cn
CD
0
*
r-
in
o
o
CO
o
cn
10
0
CO
in
o
in
O
£
o
ro
ro
o
0
p-
0
CO
0
in
r-
o
o
CD
Q
b§
o
AJ
IfJ
O
ID
1
CD
g
m
3
tM
Q
a.
UJ
—
o
ro
CO
o
o
CO
ID
O
Ol
LO
O
O
CO
o
ru
f-
o
CO
r-
o
o
p-
o
ru
in
0
a
o
1
in
o
0
in
0
3-
0
in
in
o
1
'
1
(_
o
Q
O
O
_J
CD
_l
—
0
CO
CO
o
_l
p-
CO
o
p-
in
o
o
CO
0
cn
CO
o
CD
10
o
ro
p-
0
3-
in
o
o
CO
o
i
81
o
0
CO
o
3"
0
10
in
o
CO
r-
o'
£
0
p-
p-
0
,_
0
Q
O
O
_l
CD
_l
—
O
in
CO
o
in
01
co
0
r-
in
o
cn
p-
0
i
i
ro
r-
o
s
o
p-
p-
o
'
ru
in
o
cn
o
o
o
i
CD
co
Q
p-
CO
o
co
CO
o
J_
o
o
o
o
_l
CD
-J
ro
o
^
3"
-
O
1
1
1
1
1
in
cn
o
*
ru
p-
_.
ro
ro
-
*
—
cn
ro
-
ro
ru
p-
CXJ
*
cn
p-
_J
CO
o
^
*
o
in
^
in
3"
^
^_
<
CO
X.
<
u
o
D
CO
o
f
LO
-
-
in
CO
o
o
p~
-
CO
ro
-
cn
in
—
*
—
ro
r-
_
ru
ro
ru
CO
ro
ru
3-
—
ro
p-
-
3"
in
-
CO
ru
CO
3"
_l
ifi
^
cn
CO
0
*
£
0
*
^
CO
I
<
u
<_>
D
o
o
-
o
o
-
o
o
-
o
o
-
o
o
—
o
o
-
o
o
_
o
o
_
o
0
-
0
0
—
o
0
—
o
0
-
0
o
—
o
o
„
o
^
o
0
__I
o
0
_!
^.
cn
or
u
Q_
<
cn
_j
_l
*^
z
<
u
JL
co
f
O
CO
ro
o
ru
m
o
cn
ru
0
01
ro
o
3"
3-
O
in
to
0
ro
10
0
10
in
o
ro
3"
O
CO
LO
O
O
CO
o
CO
p-
o
in
3"
0
ro
o
ru
ru
o
in
ro
o
z
o
f—
<
>
LJ
Q
Q
a:
<
Q
Z
<
t—
in
ii.o3-cooior-cor-ru
Uuinrucnoom — tnio
Ljp~coincor~r^p^coio
oooooooooo
tr
O
o
o~~J;'*!CT!^l^0.c!
inmcoor-r03-iocor-
n — rufo — — — — roin
cc
ccruoinco — ruin — co
oroLOo—-cnco — oooo
or — roru — — — rururo
cc
LJOOOOOOOOO
Ljoruro — 3- — roruco
cururop-cnoiooop-
oincoinincop^p- 3-ru
_i
coo — ooooo — —
cn in — o
co — ro in
— ru ro 3-
— ronj — —
o_ — oiinro3-mo — ru
ijoiru — coco — — — ru
u
j- GO
ro
I I i I O O O <
Ulb-lj-CnuOOOD
z: a a a in in
03 03 03 03
DI3DZ1OOO
i i
o o
o a D o Q _i
cn
CJQ_0_Q_CLCDCOCDOCj
^ UJ LJ CJ Lu _l
UJ
-117-
-------
V
<
r
in
8
oc
UJ
OL
x.
UJ
oc
1
UJ
O
OC
in
I
g
ID r- ro — 3- CD in ID
— 3- oo cn in 3~ ro o
oo — — in ro 3- r-
in in P~ in o — —
in oiiCD3-op-ini —
— 01 ruroinr-ro in
01 — 01 ru ID ro o
01
in
r-
r-
ru o p- p- oo
o in in ID CD ru t P-
— to p- ro p- in in
— 01 oi — ro 3- o
p- p- in o — ru cn
oo o r* oo o oo
in ru in 3- ru ru
— ro oo o cn 3-
ID P- P* ID ru ro
r- ru — — ID 3-
ro p- in 3- in —
— p- ID ID ro in
in in ID 3- o o
in
8i
in
o
ro
01
ro
o o ru 3- ro i CD oo
o ro ro 10 — p- ro
co — ro o ID 01 01
ru ru ID ru ru ro 3-
ioioruroincDCD
iDinror^rooO)
omr^ro — 3-3-
01 ID
o r~
ro _ _. _ _
mrjoinr-o — roru
u3ru3-roSoo — 3-F
ID ro ro co r.
ro oo
in ru
3- oo
P- 3-
>3-at~-o
in 10
ru 3-
ru o
ID ID
cnr-P* — 3-rur-rum
ruo3rororuro3-ruru
p-r~ODr-ruromior-
i ocoinr-CDOiroru
cno — ooi3-oiro
r-ooinrur- — rocn
r-p-m — — ru3-3-
oo r-
ru 10
u3 r*
p- cn
— in CD 3- o t-
3- in P- ID 01 in
ID ro ro cn in co
ID ro ro in in in
ID o — ID — —
p- Ol CD 3- P- P-
in — in 01 ro 01
in CD p- 3- — —
o ro
p- —
in in
ID 3-
SSK
— in ru
ID p* ID
O CD ID ID
cn 01 in in
— p- o ID
in r- co 3-
r- 10 3- 01
to o 01 ro
ru CD ID in
in ru — 3-
3- p- in ru
ru in ru ru
— ID — ro
— in 3- ru
to ID 3- ru
cn co ID 3-
o ru ru —
— ru 3- in
ro ru t ru
gro —
ro in
o ro ru
in
ui
oc
I I I I O O O <
£ £ 5 Q 5 in in
>•>>-> _l _J
CO 03 03 03 ~ —
D3DOOOOII
QQQO_I_I_JUO
0-0.0.0.03030300
UUIUJIJ_I_J_J3D
arzoco-Jino —
Ol
r-
(D
r-
CD r-
ru r-
— ID
co ru
ro ro
in ru
o —
ro in
ru CD
in ru
— ro
01 ru
3- —
— CD
r- ID
ID ru
CD 01
o r-
3- ru
ID 3-
o ru
ID —
r- —
ID ro
— o
ru —
00 3"
ro p-
— o
ID —
3- ro
CO O
ro
ro
£"
Ol O)
O) (D
OD ro
ro o>
pg
cn ro
o ro
01 —
in co
r- ru
3- ro
o 01
ID 3-
r- 3-
— r-
— in
ro o
r- ro
ru in
r- o
oo in
in 3-
2
— O
in —
cc »-
Ud <
< o
in
o
_i o:
2 <
< t-
uj in
i:
ID
in
in co
r- ru
01 ro
o —
01 co
o ID
ru r-
00 ro
CD ro
o m
ru in
03 ro
x t-
ro oo
ro CD
ID ru
00 3-
o 3-
ID 3-
t- CD
Ol —
ro in
in ro
ID
oo ru
ru cn
ru ro
r~ ro
3- in
3- ru
3- ru
03 ro
3- O
(O —
ro cn
r- ru
01 o
ro o
01 co
in ro
Oi 5
Sf">-
nj
in ru
CD —
ro 01
in ro
2
O
o uj
- o
_j
- o
i: oc
• <
Q O
- 2
2 <
< t-
uj in
r
s
O
<
ID
Q
O
cc
LJ
Q.
coin — roinruP-inun
ruro — — — — ruro
IDOP-O — cnoor-ru
orunj — ru — ru — ru
ooooooooo
r.-oODinoirocDr-ro
r-CDOor-roinioiDio
oooo — — — oo
ru 3- 01 oo r- — 3- \ in
r-lDODlDru3-3- Ol
oooo — — — o
* *
3- i ro 3- ru in o i —
co cn Co ro ro 3- ro
o o o — — — o
* # #
1 03 P- in — 3- (D I O)
r- CD ID — ru ro cn
o o o — — — o
« * * *
ru ID ru — ru oo i i o
r- OD cn r- in in r-
OOOO — — O
* *
r- ID 01 01 3- r- i i 01
p- CD 01 ID in in in
oooo — — o
» * *
i 3- ro oi 03 r- i ru r-
3- ro 01 CD cn ID r-
o o o — — o o
* *
i cnoiru3-roinor-
r- — CD — muDinro
o — o — .00
* * *
P- — ru;noo — ID — o
p-ioioooroioxcotn
oooo — — — oo
* * *
i i — 01 i i i — cn
ru r- o oi
— o — o
» • * *
r~rocop-oi — roioiD
oooo — — — oo
* * * * *
i ininr-r-3-oo3-o
cncniDro3-3-iniD
000 — — — OO
• *
coo i CD r- r- ro p- o
r-o lOroroioiniD
CD— O — — — O O
* *
3-CDoinoco i r-cn
P- o o ID in in coin
o — — o — — oo
* * •
i — o ru i ID ID 3- —
oo o CD r o in ro
o — o — ru o o
* » *
— P-O3- i ru CD CD o
p-o — ID in ID in p-
o — — o — — oo
* *
o 03 cn p- t r- p* i ro
cn 3- ru r- CD ru 3-
oooo — ru o
* *
< £ a. in i— >->->-
1 1 1 1 O O O <
i/iu.u.inoooor>
01-1-0 — — — zu.
£ r o Q Q in in
03 03 03 00 — —
rsir-jisitMooo<<
QQQQ_J_I_)OO
Q_Q_Q_CLCDCDCDOO
o ro
0 3-
— 0
o ro
o ro
— o
o 01
O M
— 0
o m
O PJ
— 0
0 00
o ro
~ 0
O 0
o 3-
— o
o r-
O ID
— o
O ID
O 3-
— O
o 01
O 3-
— O
O CD
O —
— O
o in
o ro
— o
o oi
o ro
— 0
o —
O 3-
— CD
0 P-
o ro
— o
o o)
o in
— o
o ru
0 3-
— 0
o p-
o r-
— 0
5°
cr i-
u <
0, >
£ UJ
< o
in
o
_J OC
z <
< i—
UJ U1
T.
U_ — ODt-ruCD3-— (DOO
u.
ui
o
ct
o
u
033-ruruoruCDooi
oioococooioirjor-3-
ooooooooo
rooiinoi — loroo —
r-r-3-o — njuoruiD
mr03-inoDOi3-roiD
.........
3-oiin3-
cDru033-
3-roior-
O)3-inroru
3-— co — 01
ocr-3-m3-iD — r-ooiD
oinoioicor~ro — t- —
oc o — — o ru — ru
oc
UJOOOOOOOOO
Ljoru3-roiDo — ruro
ocooop-ru — ruins-
rrcoru3-rococDiDro —
oiruro — inorooi3-
r-r-roiDiDio — ruru
3-3-iniD3-ocDror~
t— coro — iniD — — oofxj
Q.rooioroooi3-3-r-
uiP-iDP-rorvjiDCOinro
ccnjp-iriiniriinioinp-
Ljr-3-o — incoruiDo
i— ro OD o rop*rooro
2 i — — — ruro — —
— i i
-
l I I I O O O <
inu.u.inoooDO
oi-i-o — —. — zu.
s: £ o o o in ui
tsl
>•>->>- —I _J
in < < < < _i _i
> 03 CO 03 CO — —
2MNMMOOO<<
OOODO_J_J_I(JO
< < < <
UJ 0. 0_ Q. 0. 00 03
CC. UJ UJ U U
o
UJ
O O
O U
-118-
-------
p-
r-
cn
CO
i
>-
<.
z:
ID
8
PER 1 01
tr
UJ
u
*
CD
D
(J
-^
t/1
r
<
CC
1
<
z
<
<_>
_J
<
f-
o
fe
>
Q
3
in
1
CC
<
Q-
H
O
o
or
Ld
^Z.
s
o
>
<
ID
in
j
ro
ru
-
o
cn
CO
CD
in
j-
fO
ru
~
Q
O
CC
UJ
Q_
l/l
r>
l/l
OEZCCU-JC/lOI — ~) *
oooorur-ru i ro 3- — ru
— j- — j- in to in j~ — o in
to cn ro j- to o o — cn cn r-
— — ru — AJ ro ro oj — — —
oj o co m — r- cnincnco
oj cn — in P- ID — inruro
oj ru ru ro ro to oj — roro
I O O O ID - ruj-
I I I I O O O < U. - •
Q^^Q^H^i-i^Ll'tnc/!
>->>->• _)_J_JZZ
<<<< _I-J_IOO
CO CO CD CD — "- — (/)(/!
DIS^DOOOXXIOO
OQQD°°°3cjUXX
UJUul£j535DDDLCLC
in
CO
CD
1
ru ru
ru
CD —
Sp-
ru cn
ro 10
ru
ro CO
CO CO
ru
CO ID
o m
ru
ro o
ru —
tD CD
O ID
o to
ru
— in
ro in
CD CO
cn CD
in o
ro —
in cn
o —
r- o
Oj
r- —
P- Cl
r- o
J- ID
o o
o j-
M cn
ru
co tn
fo ru
z
ct 1^
LJ <
P
Q
2 <
UJ (J1
P- o
in co
o tn
ru
CO (D
— O
P- o
ru —
o cn
— CO
j- in
ru
ro ro
SCO
r- —
r- ru
o r-
ru
CD ro
ru to
ru -
in r-
03 CD
ffi ;e
co ru
10 —
j- in
cn fo
o CD
ro cn
CO CO
to ro
to cn
ru
o o
ro ru
CD J
ro —
r- o
ru —
01 o
ro o
10 ro
ru —
tD ro
0 CO
ru
p- —
ru o
ru —
z
~ ~
— —
O UJ
• Q
r- <
Q Q
Z <
< t-
U! I/)
a
o —
o — njo — — njru — — KI —
OOOOOOOOOOOO
r^oor-roj-inr-r-co— -o
orzoru-jc^ox — ~i^:
mop-p-roru— i mcncD —
p-cnocoruj-j- cncococo
o o — o — — — oooo
• • * •
p- i — iDj-mco ' r-j-mco
r- oP-rururu r- in — —
o _ o — — — oo — —
i P- ro r- cn — — i p-cncop-
cn o to — j- m mcDcoffl
o — o — — — oooo
o — to in — to t i — ro ru tn
0 ~ — 0 00
.rinocn — j- i i — co CD ru
r- Q — co in in cococncn
o — — o — — oooo
* * *
i — ro in co j~ * iDcnco — P~
CD co co co P~ inintoruru
o o o — — o o o — —
i p-cnrur-p-op- — oinr-
— ojco — inroincomop-
*
ro — OCOP-CO — inroruiop-
cococotOoruj'tDiOtoru —
OOOO — — — OOOOJ —
i tcooi i i — ru cn p- in
Ki co ro ID en cn cn
— o —oooo
P-OOP- — roiniDCDP-roru
o~— o — — — ooo — —
* *
i ruococniorocor-P-— o
ru — coruxintDiniocDCT)
— _o — — — OOOOO
— CD i p-p-coinroruru — cn
p-o toruruininr-cOforu
o— o — — — ooo — —
*
coruocDj-in cor-mcnru
o — _ o — — oooo —
iiniDJ"'tocncDrooi i
oor~ rocoj'Cnin
— — o — — ooo
OTCoru i ruoj-p-j- i i
r- ru — CD inr-^toco
o — — o __ooo
— cop-ini— njiotoi i
co co io LD coo coin
oooo — oj oo
<£Q_i/i»->-t->- ruj-
I I i I 0 O O < U,
uiu.u.cnouoo_jD
01— t-a — — — izt-u-Lntn
X ZrOQQ(/)t-(/)_/_;
>>>V _l_J_)ZZ
<<<< _J_(_JOO
CD CD CO CD — "- — i/lui
DDZJ^OOOXXXQO
isirsirsirvjOOO<< ruj-
i i i I O O O < U.
>->>->- _J_J_JZZ
OQOQQ_1_J_JOCJ(_)XX
C/l
ujD.o-0-Q,CDcacoc_)(_)c_)uitn
^ULdLdU)_'_l^Z»Z)Du.Lt.
Q r z or o _J i
-119-
-------
r-
01
00
x
ID
l
in
Q
o
£
tr
o
>/CUB
kNOGRAMS
z
•-
o
in
b-
o
Q
=>
3MPARISON S
0
K
t—
Z
8
o
UD
in
3-
ro
ru
-
0
en
CD
r-
ID
in
ro
ru
—
Q
O
a
in
in
bj
ct
Qrzrro_jino —
oo 00 3- ro r- ID ro i 10
3- ru o ru ro 3- ru CD
cjDCooococooiro i in
CO CO O 00 00 3" CD ID
ro in in — — — — ro
i — 01 3- oi in in i r~
ro 3- ru 3- in 3- ru
r- on CD o 3- ro i i m
o en — r~ a o r-
— in in ru oo CD i \ 3
— ro — r- oo ID in
oirorocorororuroru
r- CD co r- 01 01 ro
oococorurocDroj-o
, _ CD 3" 3- 3- 3" 03
ruroruoi — oooj-oo
in i — r- i i i 3- r-
ru 3- — 3-3-
ruocDr-cDooruoo
cDroooocooooinj"
i j- — inor~ — — ID
ocoruoooru — mo
or^cnruoocojrco
co o co r- in r~ i r~ ru
ru r- ro CD oo co oiru
ru in 3- — j- ro
ro ru — 3- i to CD o r-
ru r-- CD ru ru — ro ro
— r~ — • ro i in co to --
— inoij- 00103-
ru 3- ro — ru — 3- ru
ruojcDincor-rucD
1 l 1 1 O O O <
l/)ti.L.U>(JUUOD
Q(— H-Q — — — EU-
3Z t Q O O l/l 1/1
iiiiooo;;
OQOO_i_J_JC_)O
Q.Q.Q.CLQJCQQQCJO
oi:zrrc_>_jc/io —
00
in
in
o
r~ 3-
Ol CO
CD CD
0 CD
in co
ro in
ID —
OJ 3-
in ru
(D ro
co in
in CD
-,
0 CO
in —
3 3
ru r-
J- CO
3- CD
CD r~
ru in
ro —
in CD
ro ru
ru CD
CO —
ru —
r- —
Z ME AN (ALL SAMPLERS) I
STANDARD DEVIATION i
ru ro
ru CO
o ro
in co
at oo
ro CD
ru oo
r- CD
r- oo
CD —
o ru
CO ro
— ru
0 J
o ru
CO CD
in 3-
ro —
CO Ol
r- —
r- ro
Ol Ol
r- ro
Ol CD
ro in
CD ID
ro —
CD —
3 ru
3 en
o —
ro —
ro o
3- —
X MEANID.M.L , G, I ) I
STANDARD DEVIATION i
>
a
in
o
CD
in
ro
ru
-
o
Ol
CO
r-
CD
in
3-
ro
ru
—
Q
O
rr
bj
Q.
rs
0
H-
O
<
rr
Cnroino3-j r — CD
— ruruooooru —
ooooooooo
— 3-MlDO — CD — O
oococDiniDcDinroo
0 — — OOOO — —
ior-3-roo3-ro i 3-
ruiDinior-r^io co
r-ruiDrurucnro i en
o — — oooo —
*
i r~ in co — ro co i r~
in CD in r- r- ID o
— — oooo —
r~ en 3- oo CD r~ i i r-
co ID r~ 3- in in o
0 — — 0 0 O —
*
o in oo in ru ro i i r-
CD CD in ID r~ ID co
o — — o o o o
cnroj- — ininocoj-
r~ruoiininincDcocn
oru — oooooo
*
rococo — r (t> 3- — CD
cocniniDinminroo
*
OO3-CXJ — coroinco
o — — oooo — —
M i co ru i i i in ru
0 — O
incncooiror^— co —
r-ocninintninrur~
oru— -oooo — o
i — ruocD — rocDru
— — oooo — o
coco — or^cDrorocD
locooioj-j-j-ruru
n' _
o '^ < in — r- — i roco
r-r-roinincD men
o — — o o c — o
*
3-CDr~CD 1 J- — 3- O
oru — o oooo
• *
inruco— i ro — 3-iD
r^iDrom r-r~3-oo
o — — o oo — o
rururo — aroinro —
r^ruoininininooi
oruruoooo — o
* *
^ n Ct. in i — *~ ' — ^~
1 1 l 1 O O O <
Ulb-b-C/lOOOOD
QI-I-Q — £ U.
•£. z o Q Q in ai
>->->- X -J _l
< < < < _J _l
CD GO CO CO — —
DDDDOOOII
tsirMrslMOOO<<
QQQO—J-J—tOO
CLQ-O-O-OOCDCOtJlJ
LdUluJUJ_l_J_/DD
o ru
o in
— o
o ro
0 3-
— 0
o m
o in
— o
0 3-
0 3-
— o
o ro
o in
— 0
0 3-
0 3"
— o
0 3-
0 CD
— 0
o ru
o in
o —
0 3-
— 0
O 0
— 0
o ro
O CD
— o
0 3-
o in
— 0
o in
0 CD
'
0 00
0 3-
— o
O CD
— 0
o ru
0 3-
— o
o r^
o ID
ME AN (ALL SAMPLERS) 1
STANDARD DEVIATION 0
u_r~— oo — ruino3-ro
u-3-inoo — oo — 3-coro
oooooooooo
rr
o
• rurur~rucooinrooo
uoo — in3-rur~r-in3-
• rooru3-3-oooim
c/irocDoo — — rurucoin
rr
oru — cnininr~r~cocn
rr — ruruooooro —
UJOOOOOOOOO
ujror^ru — • — 3- ru ru ro
o 3~ ro • j" ro ni rt co
c/ioruruooooo —
rrorucDcDoinro3-r-
ocnj-ruroro — oo — oo
rrru3-cncnrooi3-ruo
ujocDoinconjroco —
3-cocn — — rururucD
D_roinj-cD'— r^roru3"
Ljo3-romrococn — co
£5 ;?;£?££ ESS
REGRESSION VS 2 INI
D EPA DZUBAY MDS-A
M EPA DZUBAY TF-M -
N EPA DZUBAY TF-P -
R EPA DZUBAY MDS-S
C LBL LOO DICOT
L LBL LOO DICOT
S LBL LOO DICOT
G UCD CAM ILL SMDAY
I UCD C AH ILL SFU
-120-
-------
oitrifiinrur-toj-
— ru — — — — MOI
r-
5
CO
1
XI
CO
cc
_.
cc.
UJ
5
CJ
5
~>
o
i n
1
rf
£5
Z
2
..
^
O
LJ
O
5
_j
fe
._
Q
t—
tn
o
£
31 Ul
< Q
_l CC
< a
~ z
z <
< i-
u tn
Nl
o 01
CO CD
to in
ru —
ru ru
ru —
— CO
o —
M r-
ru
O M
01 CO
ro o
ru —
ru o
01 in
O) ~~*
— 3
r- o
CO 3
r- cn
oo to
tn o
M O
ru M
3- ru
cn 01
CD M
—
01 ru
M —
ru —
o r~
ro 3
ru —
ru to
ru cn
O M
M *-«
m in
ru cn
ru —
ru co
— ru
ru —
to M
r- oo
—
0 0
tr 3
CO —
_~ —
z
o
t—
Z -
• >
O LJ
• Q
_J
• O
r cc
Q Q
- z
z <
< h-
LJ cn
X
Q
cn
o
CD
in
j-
M
ru
-
o
—
cn
CD
r
CD
in
3-
M
ru
„
n
o
CC
LJ
Q_
N
O
O
cc
tn
0
o
o
r-
0
o
18
o
m
0
1
s
o
r>
CO
0
i
,
to
to
o
1
3
0
1
J-
0
8
o
1
to
CO
o
0
CO
0
<
I
tn
£
^
CD
NJ
Q
0.
LJ
Q
r* M cn
— ru o
O O 0
CO
x 01 r-
o o o
r z cc
cn oo f
r- 01 to
o o o
I 3 X
01 r-
o o
ru x -
CO CD tO
0 0 O
*
— co r-
01 01 to
o o o
* •
CO J- J
X Ol CO
0 0 O
x ru M
in j- 01
o o o
0 M —
o — x
* •
nj nj o
r- i"- co
o o o
i in r-
nj r-
— o
— m ro
cn cn r>
000
* *
CD r- ro
o Ol CD
— O 0
* *
O 1 —
o r~
7 °
3 J" J-
— 0 CD
— — o
* *
Ol CD CD
O O O
* *
M cn —
— o to
— — o
* *
to j- to
CO 3 CO
000
r: Q. cn
i i i
u-u.cn
t- I- Q
j:
^ ^ *t
CO X CD
3 D D
NJ NJ NJ
Q O Q
CL a a.
LJ LJ LJ
r z cc
- 01
ru —
o o
01 r>
M m
M ru
ru j
- cn
M M
ru j-
— M
ru X
to in
Oi ru
J- CO
o cn
Ol Ol
— —
— to
ru co
to r-
M CO
— —
i i
CD —
ru 3-
— _
r- in
ru 3-
— _
0 —
J- J-
cn to
cn co
— _^
I to
j
1 tO
in
" -i
1 JT
^
—
i— *—
0 0
CJ O
0 Q
O O
0 O
_l _l
CO CD
-J _l
CJ _l
01
ru
o
X
co
tn
M
r
r-
M
J-
i
i
'
in
M
Ol
X
_
i
r~
CO
„
ro
in
„
M
r-
i
01
ru
r-
—
o
ru
ru
h-
O
o
o
o
_J
CO
_J
en
ru in
ru —
o o
o —
to to
o o
o —
i in
X
0
*
i co
i-
o
I Ol
r-
0
I X
co
o
1 CD
in
0
o o
j- o
3 3
CO ro
to in
o o
J- 3-
— 00
— O
« •
3 Ol
to r-
0 O
Ol O
J- CO
o o
CO CD
3 in
0 0
r~ o
r- in
0 0
— r
3 3
O O
o X
CD in
0 0
1 O
in
o
>_
«{
D D
n u.
cn cn
_j _j
I I
o o
Q a
O CJ
D D
CD —
o in
0 3
— 0
o ru
O M
— o
o co
0 M
— o
o —
0 M
— O
o o
O 3
— 0
o r-
0 3
— O
o r~
O CD
— 0
O M
o j-
O 0
o in
— o
o ro
. .
— 0
o r~
O M
— o
0 0
O 3-
— 0
O CO
O J
- 0
0 3
O 3
— O
O 0
— o
o r-
O 3
— o
o Ol
O CD
— 0
z
— o
cn —
CJ <
CL >
£ LJ
< D
_j rr
ro
LJ O
in —
r —
cc
or cn
o j-
U) O
ui to
CL in
o r-
cn o
cc —
o r~
cc x
u r-
0
—
i- cn
CL 0
ui x
cc r~
UJ O
t— —
Z 1
<
1
tn
z
NJ
X
CO <
> X
Z NJ
O Q
cn
tn <
LJ CL
CC LJ
co
LJ
cc a
M tD O
— cn ro
cji r~ oi
000
M M X
0 M 0
x in ru
CD M O
3 in 01
M J- —
co 3 ru
M o r~
— ru o
o o o
Ol CD —
3 CD X
o 01 co
— 0 0
M 3 f
— j- ro
o CD —
01 — r-
— in CD
M J- —
ru — x
— CO CD
— x r-
M r- ru
— 01 to
M 1
s: CL tn
u-u.cn
t— h- O
>- > v
ili
NJ Nl NJ
O Q O
Q. CL 0-
LJ LJ LJ
r z cc
— M
— ru
Ol Ol
o o
in 01
ru M
— —
r~ 3
r~ r-
f M
X J-
0 0
ru r-
ru to
M ru
— M
— ru
o Ol
in M
in 3
3- M
M in
ru 3-
3 cn
— ru
3 in
— to
t— h—
O O
CJ CJ
Q Q
O O
0 0
_J _J
_J -I
CD CD
O _l
in ru co
o j- x
cn co r^
000
0 0 M
X 3 3
oo r
ru x to
r~ cn M
CD M M
in 3 co
co tn ru
ru
o o o
in in ru
cn CD o
to ro co
— o o
o in r~
J M X
ro o j-
— f 3
f CD Ol
CD M CU
CD cn cn
— X 0
to — ru
— ox
CO CD —
1 3
J_ ^
o <
cj a 3
*— zr u.
o cn cn
_) )
_J )
Oil
o < <
-J CJ CJ
_l Q Q
CD CJ CJ
_l D D
tn o —
-121-
-------
rcoorxjin — Mt-j-roinoi
rurxifo — — — — nj — M ru nj
Ol Cft P> O O1 (D — ID O> J- J- r*
njrufu — — — at — oruaioi
o
>
ID
^ £
S
>- M
< —
ID*
— ru
i _
se
E -
* .
U
3
iJ a,
¥
5
o r-
z
ID
-i in
<
0
te *
1 „
>MPAR I SON
2
£ "
INTE
SULTS PERIOD
ru — iDM r- en M i o rn M m
in M jp rj> 01 M M ru ru — f
oi i -- oo 01 ru o i — f-~ — in
r- — gj-ODOD incnp-o
i to 03 o ru — r i M— ruin
5 jp Ol f\J CO 1C (0 CD £• nJ
in r- r* ru j — i i j- — — en
— to o in o ru i looiDoo
i (go — in— i ru M o oo o
' ^fu-oiSS^~"^ru
SfiPPSStS? 'SS
' ' E 8 ' ' ' S ffi £ Si "~
BjesjasRsapB
rur\j — M M M — — — — ru
srs8s?a5ffii&g
ru CD o ru to ID i ooru — OM
(U M 00 OT in 00 00 GO nj O) M
ioruj-i— — OJ-GDI i
ru ru M in — ru —
as$~'2S'?K' '
I I i I O O O < u. • •
533DoooiiiSo
QOaa_l_J_IOOOXX
ID
f
— M
0 0
03 CO
r- £
M
en r-
— u)
s^
8S
s ?
M —
ru °°
M m
ru
S2
w -
M r-
0 01
MEAN (ALL SAMPLERS) i
STANDARD DEVIATION
ap
— O
en j-
S "
M ID
o IS
58
*7>
gs
ru
gs
in to
ru -
M5
ru —
S P
MEANIO.M, L.G.I .K) c
STANDARD DEVIATION
O
>
in
£
M
ru
-
-
en
03
r-
(D
in
j-
M
ru
-
8
cc
tf
Nl
o
o
QOOOOOOOOOOO
ruroiDajr^ruincnr-— £ f1
o — oo — — — oooo —
o S o oo at j- j- in in en o
— o — a — — — ooo —
* * * * •
Si o en — ru ru iDJ"""ru
o — a — — — oo — —
* * *
ocnr-- — MID iniooo —
— oo — — — ooo —
o — — o — — ooo —
• *
Sp-r- — — r-i 1030 — o
— OD r- in J- P- — r- o
o — oo — — o — o —
* • • •
I lOP-OlCD-" I TMlOjj1
l—MCnJDCD in ID ID — f
ooo — — ooo — —
ihruoioj-injincrir"O
o — o — — — oooo —
• * • * •
r-CDMr»icD(D''ia)j' i MOD
ruruoocoo — Mf»o3 in o
oooo — — — oo — ru
* *
i icorui i i(j) — j-r^o
— o o o — o —
* * *
inoui — f r — — \r> \r> f ID
r- — 03co — Mininm — oat
O — OO — — — OO — — —
* • * •
i j-inruoiotn — ruifiMj-
— rur-ruj-j-intDr-r- —
*
ruo i ruMcojojru— jocn
o— o — — — oooo —
SM-iooru i j-ruoDin-
M — r^j-in r-t^j-r-M
o — — Q — — oooo —
*
i— Ojri— tD^OllO1 '
— oo — — ooo
* * *
SootoiMin — j-ifji i
ruair^ intoiDcoin
o — — o — — ooo
ffiS?£'K8'P£! '
OOOO — — OO
*
OI3-
1 I 1 1 O O O < U. • •
IJ1U.L.CJ1(J<_)OQ_J3
Qt-t-a — — — r»-ut/iui
r roQQaii-in_i_i
>>>•>• _)_|JZZ
<<<< _I_J_IOO
QQOQJ_J_J(J(JUII
<<<<_J_|_JQQODD
a.O,a.Q.ODCQ03UOOlfltf
LJULdUJ_l_l_l5Z)5»U-b.
o r~
O M
— O
O O
0 M
— O
O [D
0 Ol
— O
O M
0 M
— O
o r-
o ru
— O
o —
0 M
— o
o o
o in
— o
0 M
— o
o ru
o in
— o
o r-
— 0
§s
— o
0 J-
0 0
— 0
o in
-, o
O ID
— O
0 —
0 J
— o
0 03
0 J"
MEAN (ALL SAMPLERS) 1
STANDARD DEVIATION 0
u.
0
o
>
W
i
s
*
LJ
K
UJ
tL
UJ
cr
GRESSION VS Z INI
$
0
i
*
U
0
o
c
Ol
o
o
OD
8
i
1
N
0
<
0.
UJ
O) OI — m
OOOO
M nj — fU
""""
ru ai — —
OOOO
BS£S
in co M r-
r- — m —
r- r~ jp o
03 — — in
r- OD in —
M 1 — M
£0.1/11-
1 1 1 O
u. u. m o
>->->-
N N N O
a a a _i
< < < _i
o (D o
ooo
S~~ — r^
in ru
...
— ru —
ooo
5SS
— — o
3- — 3-
01 — O
M ID f
in ru —
M 03 J-
ID f~- ID
*-*->-
O O <
o u o
oat/i
_i
_j
§Q I
O <
_| _j (j
03 2 03 01
oooo
^|Sg
«.,
o ru - -
oooo
m o 2 —
o o — —
r— oj o in
!fi o — r~
o ID r- r-
03 en OD ID
in — P- j-
to r- M CD
— to y ru
ru T
u. • •
1— U. l/l l/l
>- tn _j _j
I I O O
< < 5 D
u u i i
D Q D D
QZZQCO-lt/lOI —
:arza: \
-122-
-------
ootor- — cpro^oro
tDJ-Qi(X.j-f\jroGDLn
p-
p-
cn
CO
i
5
CO
cn
o
o
u
0.
or
CJ
r
o
CD
i
z
••
t—
_J
in
u.
o
Q
t—
DMPARISON c
u
or
\—
z
Q
CO
0
CO
in
3-
t^}
(XI
-
o
CO
CO
r-
10
in
3"
ro
ru
"
O
0
or
LJ
0_
in
t—
— >
cn
LJ
a:
orzorcj-jcno —
ru ru — co o cn ro i o
oo P- co ro ru ru ru in
— in ru 3- co ru ro i ru
inoor-LOruroro —
i ru co p- oo — — i o
nj 3- ro ro 3- 3- 3-
in 10 ro 3- ro — i t ro
co CO in ru ro 3- co
ro — o ru 3- ru i i —
ro 3- ro ro ro 3-
oo — o CD ro LO i ruin
3- co ru ro ro ro 3-
corucnr^coconj — 10
roco ro ru — ru ru o
oooop-roroo i
— — — — ro
— i ru o i i i o o
i
LOOOCDLO — r~ ro o
cn ru ru ro ru ro LO
i oo — r- cn co cn ru
p-3-rururururuin
corucn — cncninroco
LOCnruronjrurunjro
3-lOOCOLOOO 1 — —
ru p- ro ro ro in cn
ocoruin i — ro 03 ro
ro p- O 3- 1 LOCOOO
co in — 3- 3- co f
orucDcncocnr-cDo
inr~inrururururu3-
-
l l l l O O O <
oh-i-a r u.
TL H D D Q CO cn
CO CD 00 CO •— • —
D3DI3OOOII
rsjrslrslNOOO<<
QQOO_J-J_JUU
0-0-Q.Q.ODCOCrjuo
LJLJCJLJ_I_J_IDD
azzoro-Jloo —
cn
ru
3-
3"
in ro
3- ru
00 0
in ro
CO CO
ro
LO (XI
in r
3- 3-
3- ro
CO P-
co cn
ro ru
0 0
— co
3- o
3- CO
p- r-
ro —
cn ru
ro ru
3- cn
3- ru
in co
cn —
ro ru
o r-
3- —
Z ME AN (ALL SAMPLERS)
STANDARD DEVIATION
co 3-
LO ru
O CO
p- ro
ro —
— cn
co —
nj ro
co in
o (XI
cn oo
CO LO
3- ro
-L
o ru
3- ru
co co
CO 0
3" (XI
p- 00
3- ru
CO P-
in ru
co co
— CO
in
ro cn
3" —
X MEAN(D,M,L,G, I )
STANDARD DEVIATION
O
in
o
>
CO
M
(XI
-
O
cn
CO
p-
10
in
3"
ro
ru
—
Q
O
or
LJ
a.
M
o
i—
0
(£
LO
in
o
(XI
cn
o
0
5
p-
00
o
*
i
03
ru
3-
0
LO
O
O
O
0
(XI
o
0
1
3-
3-
in
in
o
CO
in
ro
LO
(XI
<
cn
Q
>-
CD
rsi
O
Q.
LJ
Q
cn in
m m
O 0
in CD
3- m
— o
r z
cn ru
in —
co ro
3- ru
co co
in ru
o —
o in
p- o
ru
cn
0 O
*
ro
3-
— O
3- CO
3- (XI
0 O
i ro
o
ro
o
OJ O
- 01
ai o
*
r- in
ro t~~
*
ro
r-
— 0
-
0 <
O Q
o in
0 I
o <
_l (_)
_i a
CD 0
_l D
10 ID
ro
CO
o
3"
P-
o
ru
01
in
o
CO
0
LO
ro
o
p-
0
CD
p-
(XI
LO
ro
(XI
3-
o
cn
r~
o
(XI
o
(XI
o
*
o
o
u
CO
I
<
(_)
Q
o
o p-
O CO
— o
0 0
o in
— o
o —
0 CO
— 0
o —
O (XI
— 0
o —
0 3-
— o
O (Jl
0 O1
— o
0 3-
o —
O 0
O 00
o —
O 0
o —
— o
o ro
o —
.
o in
O 3-
— 0
O p-
o in
o co
O CO
— 0
0 P-
— o
0 3-
O CO
— 0
o ru
0 3-
MEANIALL SAMPLERS) 1
STANDARD DEVIATION 0
u. ru
U. CO
0 0
or
0
> LO
LJ r-
• ro
cn ru
r
or
o ru
a: in
LJ O
LJ ru
a. 3-
cn o
or ro
§0
LJ 3-
(XI
1- 03
CJ CO
or o
LJ
REGRESSION VS Z INI
D EPA DZUBAY MDS-A
p- cn
3- ro
o o
ro o
in 3-
00 LO
— ru
0 3"
3- in
O O
ru 3-
3- 0
— o
cn co
co co
O CO
CO 3-
— (XI
cn ro
co co
— cn
in ro
r ex
i i
LL. U.
1— 1—
>- >
< <
co m
D D
rsj KJ
Q Q
< <
o. a_
LJ LJ
x: z
in co
CO -
o o
ru 03
ru ro
rji in
cn ru
o o
— 3-
in 3-
0 O
CO 00
in OT
0i ro
co in
co ro
CO CO
ru ru
P- CO
cn i-
I O
cn o
§5
§0
rsi o
O _J
< _J
CL CD
u _i
or o
ro
ru
o
cn
CO
r-
o
10
cn
0
3-
cn
ID
r-
03
o
ID
—
0
O
o
o
o
_l
_l
CO
_l
_l
ru ro r^
00 ID OD
O 0 0
ro o co
CO 0 0
oo ro ro
in 3-
ru 3- P-
ru ru ro
o — —
P- O P"
P- CO 3"
o ru o
0 3" CO
o cn —
ro ro cn
cn 3- ro
CO CO
00 — cn
in ID 3-
3- (XI 00
o 3- in
— in ro
o <
ODD
Q CO CO
Oil
o < <
_l O O
_l Q O
OD U O
_l D H)
CO o —
-123-
-------
i-
01
CD
i
—
5
£
to
1
~"
l/l
8
o:
UJ
Q.
tt
Ixl
J—
UJ
E
O
H
CD
13
CJ
to
£
o
§
z
..
-
u
o
3
u.
0
o
t—
01
z
o
in
cc
<
Q.
g
U
u
or
UJ
t—
z
™
—
s
0
^
ID
in
£
[2
ru
_
_*
0
on
CD
^
CO
in
3"
ro
ru
-
o
o
~~
CC
U
Q_
01
h-
_)
D
in
Ul
rr
ozzrro-Joio —
o 01 in r- r*- o o i co
r- m o m r- — — r-
m ni ro m
— — t\j ^-t ru rn ™l
CO 1 Ol J- — o [-* 1 in
ru ro ro ru in 3- co
ro ro ro 3- 3- r i
i j- o ro j" o ro i —
CO — CO CO 3" CD 3"
ru ru — ru ro ro —
o — r co — o i i CD
O ID 01 01 f~ Ol rO
ru - ro -
CD r ru ro ID o i i cn
cn ru 01 cn cn
— « (\j — •—! (\j ru
i in cn in ru in i — in
cn co — r- o co —
ru ro j- —
i inoiinin — oj-ro
3-intnoiuior~co
rinrorococnj-cn i
rurui-iDr-oooro
~~
i i in ID i i i o r-
ro (xi (xi in
ru — —
ro cn j- 3- j- — onjro
oioirucor^rur*~oj'
— ru — ru — —
i ror-o--j-oruo
oinmcoorucnro
(Xi (Xj —• (xj ro ro ~~
r~ o i (xj 3~ o — o r^
— in coiDinrooco
oj ro — ro ro j- —
corurotD — cn i r~ o
01 ID Co in ru 3- roro
— ru ru — ro ro —
i r o 01 i o co ru in
(xj (xj o ro in 01 in
ru ru ru ro 3-
— r^ in ru i — o o in
to ro 01 r~ roiooioi
— ru ru — ro ro
ru o r~ ro i o in i ro
j- 01 ro r 01 in —
— — ru ro —
<:cQ.ui»-t-i->
l i I I O O O <
U1U.U.01OOOOD
ot-i-o — x: u.
£ £ a o a in in
>_>_>_ ^_ i i
^ ^ ^ ^ i i
CD CD CD CD — —
DDDSOOOII
NNNNOOO<<
OOOO_I_I_IOO
<<<<_i_j_jaa
CUQ.CLQ.ODCDCDOO
UUUJUJ— 1— 1— J^O
ozzcEO-JUio —
o
o
-
ru
co 3-
0 CD
ru
ro ID
3- ru
in ru
in co
ru
in 3-
CD cn
(XJ
0 —
Ol CD
co cn
Ol ro
co cn
CO 00
ru o
ID ro
in 3-
ro r°~
ro in
(XI CD
ru
r- ID
— CO
ru
0 —
CD ro
ru —
r- r~
— O
ru —
r- r~
ru ro
ru —
CD 3
— o
ru —
r- 3-
CD —
z
- o
in —
CE 1-
UJ <
_i —
Q. >
£ u
< o
in
o
_J CE
_i <
< a
- z
z <
< t-
Ul 01
•£.
NI
01 ID
r- 01
CO ID
CO CO
ru —
CD 0
3- 0
ru —
r~ co
o cn
CD CD
3- in
r~ in
CO o
r o
j- ro
cn in
CD .7
ru ru
Ol 00
ru ro
CD cn
— CO
ru ru
(XI —
in —
cn ru
_ ~ _«
in CD
r~ rxj
ro ru
CD 0
cn o
in 01
-z.
0
t—
~ <
k^l _l
• >
o uj
- o
_l
• o
s a
- <
O 0
~ z
z <
< t-
Ul Ol
X
CJ
o
01
o
^
CD
in
3
ro
ru
_ .
r~
O
cn
00
•""
ID
in
t
ro
ru
-
Q
O
rr
UJ
a.
NI
0
O
CL
ruroro — — — — roro
— ruro — ru — ru — —
ooooooooo
oo — o — — — oo
orzcto-iuio —
ruiDCOtnrooioi i r*-
oooo — — — o
*
ID i 01 r* ro — o i in
01 01 01 ru ro ro ru
o o o — — — o
i f ru ru — ro ru i in
o co r- — ro r in
— o o — — — o
*
CD ai — in o r~ i i ru
r- cn — r- j- j- in
o o — o — — o
ru ru — in ID ru i i ru
ID — o ID in in in
o — — o — — o
* *
i 01 in o o r- i — cn
r 3- — 01 o j- in
o o — — ru o o
*
1 PSSSlriS?!?
*
cnoconj(Dj-coro i
ro;r — orxjj-iocO
^^ ^^
• *
i i m j- i i i cn rvj
r"" o) oo j"
— o o o
* *
tocnorui'o.Tioio.T
cooooconjj-ioj-to
00 — O — — — OO
* *
i .rcDCJ>o>or~(Xjo
cn — (jDojj-j-jrto
0 — 0 — — — OO
* *
mm i oomiocoro
COM r-.rroiof'or'O
o— o — — — oo
— . — o nj CD — i roj-
0__-0 OO
*
i cn r- nj i ID ru — j-
CT) cn cn j" o j" nj
o o o — ru o o
* * *
j-cniDcn i oj in — f
o — — o — — • o o
*
in J" fxj 10 i -T nj i co
CD in nj co r- *~ to
OOOO — Oj O
<2-aul(_h_h_>_
1 1 I I O O O <
LnL-u.i/)<_>c_>uaz>
o i- »- Q ' — s; u.
s: s: Q a a (ft ai
^_ >- >. >_ i i
^ ^ ^ ^ i i
CD CD CD CD — —
DDDDOOOII
NJNINNOOO<<
OOQO_I_)_IOO
Q_a.a.o_cbcQCDoi->
UJUJUJUJ_i_l_l=)D
otzrro—iuio —
0 CD
o r
— o
O 0
0 J-
— 0
o r-
o ro
— 0
o ru
o ro
— 0
o ui
o ro
— o
o ro
O f
— o
o —
o r^
— 0
0 t-
o r
~~
0 CD
O J-
'
"
o in
°. "!
— 0
o oo
o ro
— o
0 0
0 3-
— o
o o
o in
— o
o 01
0 J-
— o
o o
o (D
— 0
0 00
0 f
— 0
0 CD
o to
— 0
z
- o
01 —
CE t-
UJ <
-I —
a. >
^ UJ
< o
01
o
_l CE
_l <
< a
- z
z <
< t-
LJ 01
u.
u.
o
CJ
CE
o
o
>
o
^
cc
Q-
o
rr
oz
UJ
UJ
CL
o
01
cc
0
a
u-
ut
»—
Q_
Id
o
cc
Ul
z
01
Q
*-.
01
UJ
g
UJ
cc
o r*- ru
r- in r-
o o o
ru r- to
ro ro ru
— ID tD
o in to
(XJ J- ID
cn in CD
o ru ru
O 0 0
3- ~ Ol
ro in ro
o J- Ol
— — 0
in — 01
01 in in
CO 3- Ol
o ro —
ru in ID
ro t^- 01
.r in o
ID in 3*
r- in 01
r o —
i —
1
< r a.
i i i
in u. u.
o t- i-
>- >~ ^~
< < <
CDCOCD
Q Q 0
a a. a.
o r z
oo in ro
r^ ID 01
O 0 0
CD in co
3- 3- ru
ro co o
co r- r-
ru 3- ro
(D
— ru —
O 0 0
r- to 03
— 01 r~
CD — —
co in ru
0 ID CD
— 00 CO
(D cn co
oj j- ro
in ro nj
fo OT co
to r- J"
?$s
in t- t-
1 O O
01 O O
o — —
r o o
>_
<
CO
3 0 O
O _l _l
0. CD CO
rr o _i
ru 01 j-
ID ro CD
o o o
ro ro —
— o in
CD m in
— ro 3-
in ru ro
Ol CO ~~
ro ro cn
ru — —
o o o
CO o co
— in ru
ru ru —
— o o
in CD in
3- ru r~
CO CO Ol
ro r- ru
in ru 3-
ro co in
3~ CT) (D
r~ — ru
3-3-00
CO J CO
,_ ,_
0 <
U 0 D
— n u.
o 01 in
_j _j
_i _i
o i 5
_J O O
CO O O
01 O —
-124-
-------
coj-cnorucnfo — ojo — m
§
m
V
xi
CD
1
I
CC
tf
or
UJ
5
LJ
GRAMS/ Cl
|
UJ
U.
1—
o
h
1
cn
g
in
cc
1
u
z
_
g
CD
in
T
ro
OJ
-
0
cn
CD
^
ID
in
3-
to
ru
—
g
or
cn
i
or
orzoru-jcnoi — ~> x
§i oor>inin i COCDOI —
tfipjfoinm innjmj
oj nj — ru oj ru — — ro ro
i o o o — CD o i ojcncnoj
— ru — ru ru ru • — —
ooooinoi i — 3- — T
ror-3-ojcoo coooo —
— ru oj — ru ru — -^ ru ru
o o o o co — i ir-rocnin
["» O — CO O — ID3-CD —
— — _ — oj OJ — — —
i oooor* i tororocnro
— in •" oo "• coor>coco
„ — — oj nj — — — ru oj
i ooomto3- — oiincoip
oicnoroioinocorocoro
coofoco — ojcn — r>3-ru
-OJ--OJOJ -
Soooo— 3-inor>r-r^
oojr>r-cnruojoicnrocn
3pr*-iDinr-cooin3-3'OjcD
i ioo< i icnco^jcnr-
— cn 3- — cn r* in
o oj in ro to ro in
1
ooooinioojr-oin — ro
ruinr^oinrucncncDoo —
ino — in — 3-r^oforuinin
— ojoj — ojruru — — — ruoj
i ooainr^TincnnjcDro
OlCD — tO — OOJCO — O13-
oo, op^™,^^
ID ID inminofor^njiDCO
— ru — oj ru — — — ru ru
oooooro i cninmroj-
— cn op — o ro — — — — 3-
j- — cn oj j- m minotoru
— ru — — ru ru — ru
lOooicncD — coat i
cn r- — > com — com
O — CD 3-rO — IDOI
ru ru — oj ro — —
OOOOI— CO — 013-1 1
jior-o— rjDincninru
OJOO— — 3- O — O
— ru ru — ruoj
OOOOiOCOi3-(£i I
cococno coo ooj
ro 3- — oj rocn oo
njru
I I ' i O O O < U. *
O>— *— Q~"- — E^~u,LOtn
<<<< _J_J_|OO
DDD5QOOIIIQQ
lsjhjrMNOOO<<
S —
CO
p- 00 ID
r> ru in
j- cn
o to
ru
CD in
mo ro
Sit
01 ID r\J
in in
r- CD -
in ID
in in
§Q
0
* »- or
- < UJ
„ _. Q.
=•§
O Q O
- Z K
z <
< (- 0
LJ Ul —
x or
co in — r-
O — Oj O
O Q O O
0 — — 0
£03 co in
co o r-
in i r- to
OD o cc
o — o
I CO CO 3-
cn o r>
O — 0
— — cn 10
r- o o ID
£ ru o; IT;
OOOO
— — — o
in — — i
— ID o i
— — ru
CO 3* fO I
o oj ro
in 0 1 1
oj ro
o — — o — —
o — . — o
i co ro o
ooo
i nj CD CD
sss?
OOOO
•
1 ' 3* ID
ro CO
— o
co co ru r-
r- o — r-
o — — o
*
ru — r-
r- ru r-
0 ~ 0
en ro — co
r- Oj — CD
o — — o
'§~S
— — o
CD 01 r- o
r-- oj oj r>
o — — o
m cn ? r*
oooo
< £ 0- Ul
E E
iill
O Q Q Q
Si UJ & Si
Q 3; z o:
cn r> i m
3- in co
— — o
CD ro cn o
•
ID ro — oj
o oj 3- r>
_, _ _ o
i i i ro
o
— — — o
oj ro ro CD
— — j- (D
_ o
in oj i m
ro j- co
— — 0
— _ o
i r- 3- co
ro in ID
— _ o
i — ID i
in co
o o o <
o Q Q cn
_)
888i
_j _j _i u
ogcnmo
u _j in o
g^S 2
OOOO
00
r* co o o
??Sft
3- oj in m
ooo —
SS^s
o o — —
•
oj cn oj r-
cn in cn o
OOO —
SSffl?
o o — —
? ? SS
*
10 co r- ol
o o — —
OU "*. OJ 3-
CD cn cn o
O O O —
O O — —
r- in cn o
* *
— cn in o
co in ru to
• •
00-
co in cn oj
ooo —
63' '
o o
ro 3- i i
0 0
£ID I I
CO
O O
u. ™ T
>~ Ul -i _J
_1 _J Z Z
_1 _J O O
U 0 Z I
D 3 ul E
I — -3 *
— 0
0 —
o ru
0 CD
O OJ
— o
0 0
o ru
— o
o in
o ru
— O
o r*
o ru
— o
o o
o 3-
— o
O fO
- o
o to
— o
o r^
o —
— o
o o
o ro
— o
O 00
o nj
o o
o ro
— 0
o cn
o oj
— 0
0 0
O 3"
— o
O CD
o ro
— o
o in
0 3"
— 0
si
UJ <
£E
3°
< Q
— Z
LJ Ul
plf
EggSg
CJOOOOOOO
g
*"*
• CO CD -
>roor-cncoinro
U3-r>runjcococo
• m ro r- r- — co —
inoinruroj-cnin
j;— ojro — tu — ro
DciDOir-in3-cocn
or-oirocnr*roin
or o — r
U O — — OJ
UJOOOOOOO
tf_0-
o co ru -
C/l O — -
SESSS
- f> — O —
- o .
ccrofutocooocn
o — ojior-iocno
or— o — co — ro —
or
Ldcocrincoh-cotn
fO j- - CD - 3- ID
— to 3- — ro oj 3-
t- 3- ID
Q. r~ in
LJ CD cn
£ 7 oj
Z I
< =
cn u.
> m m
Z rsi N
O Q Q
Ul
Ul < <
U Q. 0.
or LJ LJ
o
CC 0 E I
— — in ro
r- - * ro
— o cn co
ru r> ru ro
co o co in
Ln t~ h- h-
1 O O O
cn o u o
r o o o
5 o o o
N O O O
Q _l _l _l
< _J _l _J
Q. QQ CD QQ
^ cc o _i in
in 01 oj 3- o
CD cn 01 ro o
CD co in co cn
o o o o o
r^ 3- cn ru in
3- in in 3- ro
co r- 3- 3- —
ro in oj oj o
oj — ru to ro
ru cn in cn ro
o o in ru —
ru ru oj
o o o o o
co r* in o j-
o fo ru in in
in co 3- — 3-
O O O — —
CD CD cn in cn
CO ID CO O —
ro — 01 CO ID
ro — oj 3- ro
oj r*- 03 cn o
ro cn ro r> 3-
3- OJ ID Ol OJ
CD CD co ao ro
cn o in r- r-
ru — ro i 3-
> OJ j-
< U-
O _l D
E t- U. l/l Ul
iiI§S
CJ U U I I
Q O Q D Z)
u u u in ui
o x — -> y
-125-
-------
iricnrocncncnCTioo
ru —
o — —
r-
r-
cn
CO
i
ID
in
Q
0
LJ
0-
cc
LJ
z:
o
m
D
O
kNOGRAM?
•z.
••
UJ
u.
cn
u.
o
1
h-
DMPARISON 5
o
or
»—
z
Q
cn
o
CD
in
3-
m
ru
-
0
cn
ro
t-
(D
in
3"
CO
ru
—
RESULTS PERIOD
QZzrro-Jino —
o 3- in OD — r^ in i ~
in r- — co r~ r~ 10 —
CD r~ 3- co ro r- — i 3-
in co o in ru o — r~
ro r~ r~- ru ru ru ru co
i — cn r- — 3- co i 3-
ro in — cn cn cn in
3- 3- nj — — — ru
— in 3- cn ru r- i i ro
co ru r- — 3- co CD
ru in — ru ro — i i —
co CD ro ro ru cn cn
oo — ocDincoiDin
— 3-rucorurucocoin
co CD ID ro co 3- — — —
roro — o3-iniororo
mru — 3-rocomr~io
r-r-ruininininai —
3- i r- 3- i i i — r-
co in co co 3-
0"irucooocor~3-o
co ru 01 co — — — COID
t rurucninrur^cnco
in to r~ — co — coo
inoco3-incnocoro
— 3- co co ro ro in — ru
ID ID — co ro — i r~co
o o ru in co 3- r~ —
coioinrururu 3-co
3-coruin i ru ru — r-
3- ro r~ co COCO3-3-
3- — — in i r~ co co co
cn — 3~co — ru co co
ru in 3- ru rurucoru
LOO — 3-mocoror~
10 — in o cn o o r- —
->-
1 1 1 1 O O O <
inu-u-inoooaiD
QI-H-Q ru.
z; 3r Q a Q in cn
m m m m — —
DDDDOOOII
CvJNNNOOO<<
DQOQ_I_J_JOO
<<<<_I_/_IOQ
CLQ-Q-LXmcDCOOO
LJLJLJLJ_I_I_ID3
arzcco-jcno —
CO
in
ID
in
cn in
co ru
o cn
01 (XI
ro nj
in CD
ro ai
ru ~
cn co
CO CO
ru 10
0 O
0 CO
r~ co
ro ai
ID ru
- 5
in —
3- —
ID CO
in o
in cn
ID (D
— ro
0 0
cn 3-
in —
in ro
ro —
CO 0
ro o
2 ME AN (ALL SAMPLERS) L
STANDARD DEVIATION i
m CD
o ro
— cn
ro 3-
3- ru
0 —
— in
ro —
ro ro
cn CD
r~ 3-
0 —
ro ro
o cn
ro r-
ro ru
ro r-
ru ro
— cn
J- —
o j-
[^ o
cn cn
— CO
CO —
ru cn
U) 0
co in
CO —
3- cn
in —
in o
ru ru
CO —
ro 3-
in o
X MEAN(D,M,L,G, 1 ) >-
STANDARD DEVIATION i
o
in
o
co
in
3-
co
ru
-
o
O)
ro
r-
co
in
3-
CO
ru
—
Q
O
CC
LJ
a.
fsi
O
t—
O
t—
CC
coinoiDiDcOLOrucr)
— — ruooooruo
ooooooooo
cnoorooco3-incoru
0 — — OOOO — 0
0103-03-10— i cn
cor-inr-coioio r-
— — — oooo o
— ru — CD r- ro 3- i CD
cnorjoioinintn cn
oru— oooo o
*
i 01 — CD r- ro o i cn
10 CD r- ID CD r- ro
— — OOOO CD
*
co — co ID ru — i i in
cn CD r~ in CD CD cn
o — — o o o o
*
co 3- co r~ f co i i ID
cn in 3- r~ r- ID cn
o — — o o o o
3-ororu — — 3-3-10
roocOCDcDlDCDorj)
oru — oooo — o
* *
omror-ioai — LOCO
cnroinr^inincoruro
'
*
cor~rucororo3-3-ro
o — — ooooo —
* *
co i ID co i i i CD in
o — o — —
*
— rocoincno — 3- —
roooicDiniDcor-o
oru — ooooo —
*
i 01 CD r^ cn 3- o 3- ro
r~3-r~iniDiDcoco
— — oooo — o
cn3-rococnoruroin
r~roroiDinioiooo
* *
ro 3- — en 3- in i cnm
roiO3-coiocD rum
0 — — 0 O O — O
oru — CD i corucor-
o — — o oooo
roruoin i cn — r~ 3-
cncD3-r- ID r- — r-
0 — — 0 00 — 0
3-ioruor^cn — CDCD
0 — — 000000
*
I 1 I 1 O O O <
QF-t-Q — — — ZIU.
£ z: Q Q Q cn in
CO CO CO CO "- • •—
^^Z3!DOOOII
rslNNNlOOO<<
QQQQ_I_I_IIJO
o-O-Q-O-Comcooo
oz:zcco_icno —
0 3"
0 3-
— o
0 CD
0 3-
— o
o cn
o in
— o
o in
o 3-
— o
o ro
o 3-
— o
0 LO
o ro
— o
o —
o in
— o
0 ID
0 3-
o cn
0 CO
— o
o ro
— o
o ro
o in
— o
0 ID
o 3-
— 0
o —
o in
o cn
o ro
— o
o CD
— o
o (D
o ro
— 0
O 10
0 3"
MEANIALL SAMPLERS) 1
STANDARD DEVIATION 0
U, 3-
0 0
CC
o
> in
in co
z:
a:
o ru
CL —
LJ O
uj ro
O CD
cn o
gr-
cn
U ID
i— —
ce co
REGRESSION VS Z INI
D EPA DZUBAY MDS-A
CO O — —
oooo
ru in cn —
in ro ru —
ru ru o o
oooo
ru 3- o 3-
— — o o
r- in r- 3-
— — r~ in
o co — o
CO O CD CO
z: 0. cn j-
1 i I O
u. b. cn o
1- (- Q —
z: a
>->->-
< < <
m CD CD
D 3 Z> 0
N M Csl O
Q O O _J
< < < _J
Q. Q_ Q. CO
LJ LJ LJ _)
r z o: o
r;
0
Q}
CD
UO
fXJ
o
o
-
r*"
0
t"
QD
Q")
cn
cn
t—
o
o
Q
O
O
_J
_l
CO
_J
_)
ro CO ro
(Jl CD IT)
000
in — co
cn — r^
ru ro ro
o co —
o o o
co in ro
00 —
CO CO CO
co — co
cn in cn
ru ru ro
cn 3~ o
3~ cn cn
— — in
— in ro
i- >
o <
o o. o
— z: u.
Q in cn
_i _i
_j _i
Oil
o < <
_l O O
_l Q Q
moo
_l D D
cn o —
-126-
-------
r~
O)
00
Z
ID
I
R; PERIODS
ui
o
m
u
kNOGRAMS
U.
UJ
o
CL
L_
o
V
s
f—
DMPARISON !
1 INTERC<
Q
CO
0
ID
in
3-
ro
ru
o
cn
CO
p-
ID
in
3-
to
ru
RESULTS PERIOD 1
o i:
3- in
3- co
03
ID
ro
o cn
p- j-
3- P-
03 ro
' ft
in
I ID
cn
" "
3- cn
£S
1 O
3-
to —
ru cn
ro co
in 03
1 03
ru
CO Ol
ru ro
— in
O ID
< DE
t 1
l/l U.
Q »-
tsl rsi
O Q
Q. Q.
LJ LJ
0 5Z
•z. cc
ID 01
p- in
O) Ol
3" —
00 P-
ID (XI
p- CO
So
co cn
7 rn
ru —
ro co
3- ru
01 in
ID Ol
01 to
Ol —
f f
'3-01
cn cn
ID ID
to 3-
ru to
I ID
in
ru
II
3- in
3- ru
en P-
CO P-
in co
in —
cn 01
ro 03
CL l/l
i i
li_ CO
I- O.
>z
ii
N ISI
0 Q
O. O.
LJ LJ
2 CC
o _i in
ru in ro
ru
3- co 3-
ru ru ru
o 3- ru
— 3-0
— oj ru
ro ro i
3- ID
ru ru
ro o i
03 ru
— ru
3- ru i
in cn
ru to
(XI (XI
r- 01 ro
co o cn
p- co 01
— ro (D
i i i
in ro in
3- — r-
o in P-
ru 03 co
fit nt fvi
p- p- ro
cn cn o
ru ru co
ru ru i
ru ru
1 P- 3"
o cn
(XI (XI
i 3- ro
ID to
O) ru
— ru
1 O O
ID 0
O (D
ru ru
t— t— t—
O 0 0
o o o
O O Q
o o o
o o o
_l -I -J
-I -I _l
CD CO 03
-J _J _l
O _J l/l
o —
1 O
i ru
in
ID
1 S
i UD
• a
p. CO
01 —
ID p-
0 J"
P- 0
ru CD
OB m
cn 01
to in
to 3-
ID Ol
ro o
ID ro
in in
O) 03
(XI (XI
3- o
CD P-
o ro
ID in
OJ —
ru cn
p- p-
1 01
o
O 13
E u.
in 01
I I
0 0
o o
o o
D Z)
o —
0
<-">
(D 00
— ro
p- ro
ru 01
Ol ID
co ro
ID to
ru 3-
P- 0
in cn
— ro
ro p-
01 —
— CO
CO cn
(XI Ol
j- o
ID fU
m ru
o ru
2- O
OJ O
a?
CD m
p- P-
3- 3"
— r-
3- in
3- ru
3- 03
3- O)
ro p-
ro in
Z ME AN (ALL SAMPLERS) 1174
STANDARD DEVIATION 889
co r-
o cn
ru 3-
ID O)
- ru
0 p-
o —
0 p-
ro ro
ru co
in p-
O) ID
0 CO
0 0
P- 0
ID (XI
01 O)
ru 01
Ol
(XI (XI
Ol ID
CO ID
10 ro
ro p-
3- CO
ID —
3- (D
ID ID
3- —
ro ID
p- cn
— p-
in o
O) 3"
— in
X MEANID.M.L.G, I ) 1084
STANDARD DEVIATION 676
>
u
a
01
o
ID
in
3"
ro
ru
0
Ol
CO
p-
ID
in
3-
ro
ru
0
O
DC
UJ
0-
rsi
O
o
f—
CC
03 3-
(XI
ID —
O (XI
a o
p- p-
P- 00
0 O
Q z:
p- p-
o o
ID I
o
I (XI
00
0
o ro
r- cn
o o
p- cn
o o
*
I CO
3"
0
1 O
Ol
ss
O O
1 1
- CO
00 DO
i ro
0
p- —
p- —
03 ru
p- —
0 —
1 3"
o
— U3
P- —
O —
0 0
< >z
1 1
l/l b.
Cl >-
03 CD
rsi rsi
a o
CL Q.
LJ LJ
Q H
— 3"
ro —
Ol —
(XI -
o o
in 01
O) P-
CD O
Z CC
p- in
O) P-
o o
*
Ol 00
0 O
in —
cn p-
o o
01 ru
Ol P-
o o
*
ru m
— io
— 0
ID P-
ro o
o —
*
58
o o
00 —
— o
ID ro
Ol 00
*
(XI —
0 p-
*
1 3"
r-
3" 03
O ID
— 0
O 3"
— O
*
cn ID
— o
f ID
ro r-
0 0
CL l/l
1 1
b. l/l
t- O
z:
03 03
N fsj
Q Q
CL CL
LJ LJ
Z CC
ID ru
(XI 33
ru —
0 0
ro in
u _i
ro —
(XI 3
in ru
ru ro
01 01
0 (XI
03 ro
3- in
— 3
in in
O) —
CO O
— ru
(XI —
— in
to cn
ro m
i <
OD in
ru T
ro a-
to r
3- 3"
ro ro
ro ru
in ID
i t
i p-
3-
i in
r~
o o
o o
5 5
0 0
0 0
_) _J
03 CO
O _J
ID ro p-
— ru ru
03 3- in
ru — -
o o o
p- Ol dD
ID in in
— o o
in o —
ru i 03
3- r~
— 0
P- 1 3"
ro ro
— 0
CO i p-
to p-
— o
i i r-
ID
a
1 1 O
in
o
I 03 0
in ID
o o
co ro CO
in 3- ro
P- — —
00 P- ID
— CD 0
i ID in
o o
cn in in
io in in
0 ID —
in 3- in
3- 03 o
03 in m
O D
1 3- 0
p- in
0 0
03 P- in
nj o o
f- 3- cn
03 in in
— o o
— i ru
ru in
ru o
i- >-
o <
u o D
— r b_
Q in 01
Oil
o < <
_l U O
_J Q Q
CO O O
_l O D
l/l O —
o ro
0 3"
— 0
o —
o ro
— o
o 03
o ro
— o
o 03
0 (XI
— o
0 03
CD tO
— o
o —
0 3"
— o
0 Ol
0 03
— o
o m
O 3-
0 p-
0 3"
— O
O U3
— O
0 Ol
o ro
0 —
0 3-
CD ru
0 3-
1 — '
o —
0 3-
— o
o cn
— 0
o ro
0 3-
— o
MEAN (ALL SAMPLERS) 1.00
STANDARD DEVIATION 0.71
b_
U.
n
g
#
cn
i
n
LJ
LJ
Q.
o
I
LJ
I—
OL
Ul
REGRESSION VS Z INTERC
03 —
O3 ro
0 O
3- 0
P- 03
ro CD
o Ol
o) in
ru
o ru
o o
CO (U
to cn
03 —
to m
P- Ol
ru o
0 3-
— ru
— to
o o
to ro
o CD
1 3-
< £
1 1
in b,
D <-
r
CO 03
N r-i
Q Q
o. a.
LJ LJ
Q r
- Ol
in 3-
0 0
3" 3-
01 in
r- 3-
3- (XI
3- in
to —
ru —
CD 0
to 01
ro o
— P-
CO CO
o ro
ru 3-
CO CO
CO 03
ro «
cn co
P- (XI
ru ro
0 0
ro o
ru —
i
0. Ol
i i
b. Ol
1- Q
03 CD
D ~3
tsl fsj
Q Q
CL CL
Ul Ul
•Z. CL
ro —
ro p-
0 0
3- ro
ru U3
co ru
(XI 03
cn —
(XI (XI
*_i Kl
ru —
O CD
ru cn
3- o
03 r~
ru ru
P- Ol
3- o
in in
ro ru
03 03
o in
ru co
0 0
o 3-
ro in
h— f—
O 0
(_> (_>
Q Q
O O
O O
-J _l
CO CO
U _l
co o ro
on P- 3-
o o o
3- in 3-
co o 3-
ro ro —
Ol (XI —
in ID ru
to — ru
to — —
O CD O
(XI 3" 3"
01 cn —
— IO 3-
ro ru —
03 03 to
o to ro
Ol 00 3-
ro o 3-
3- ru ru
o r~ 01
on r^ 03
in to 3-
cn ro 03
ru ro cn
ID ru —
u a o
— z; u.
Q l/l l/l
_! _J
Oil
o < <
_l O O
-100
LO u O
_l D D
O) O —
-127-
-------
(D minjin — — miiniinin j ro cn— ID
— ruiDiocnruiDro j~ ID ID coro r-ro —
i in ID in in m — t o — M —
— ru — o ID — co — men
ru ru — ro fu M — ru flj fu
j- cn co — M m i i en o r- co ru ID
CD ru M m ru in OOJIDOI ID "i
i cn r- 10 — CD i coco i inin
nj M to — ru — nj
i iDror--, coj-ruruiD i o —
<\j — roio — in oj — MM
co — — cojocnr-ro i r- cn
cnru — ruj- — M — — — ru
oooooooooooo
ru — r-r-cor-coiDrucnruiD oin
uDocooinr^iniDincnr-ai o j-
— _o — — o — ooooo — o
a E Z CC U _1 Ul O I — ~> X.
M i f— r~- o cn
in r- t1- OKI
— o o — — o — o oo — o
otK)r-vOKiiD\oiro.K) ot~-
ro CD d ru ID ro j- CD'D o j-
„ „ o — o — o oo — o
o o o — — — oo — — — o
— o o o — o
O CO
O M
— o
o ru
o ru
— o
— ro co cn M i
— O O O — I
i in CD M ru
— o — ru o — o oo — o
oo— ruo — oo o— — o
it cn in co in ro r- i
L. r^ co CD 3- cn cn i
LJ cn cn cn cn cn cn <
o • • • • •
(JOOOOOOOOOOOO
(j_ UJ U' UU U» r-j i~- iu i^ W — i
L.t^cocDjcncncncncn cocn
LjcncncncTicncncncncn cncn
o • • • • • • • •
cn o — r-
j-CD
aij-
MOOO —
j- — r- j- i r- co
M ro in j" in cn
o — o o o o
o
o o o o — o
cnj-j-conjiDr-mcxi co.r
>roj"j~innjcoiDcnco Min
Q . . ......
• — omoiDcnocnj- orufxi
Lorurj — ro— — ru —
cn ID cn
in ro ru
j ro
in ID
S¥
m o
^s
ru m
ru in
— — ru — ro
in — ru
— ru — — M
in — M
o co cn cn r-
ro — ru —
j- ru ro —
M ~ ro ru
j- cn ID
cn — cn
— o
— j-mcoMMr-M
0 — 00 — 0
— ruo— ooc^o
rxj-iDinMj-ini-
(O
r-
o
o
0 CD
o in
— o
O LO
— o
0
cc
or
0.
o
ooooooooo
mocnco — — rucnr-
cn ru
o o
r- M
cn M
ID cn j- (
cn ID nj !
• f\J — CO M
| — — OOOO— — O
LOO — OO — — — OOOO —
OlQCDM — rUJ-J-CDID o —
CKMMJ- — j- — inioiD j-in
or • • ....
uicnr-inoLDMj-M — OCDJ
r-ru — nj — in co r- cn — i
CD aj o o
ifij-j — i r- (D ru — r~
j-ocno cnrur-iDr-
— — o — o — o o o
cn j- ru ru i — r
o ru
— o
(— — OinCDCDr~OOlfv' MCD
O-rocoruj-coiDror-id — r-
ujcncDrucotDioMinr- j-o
o
oroj"rucnotDin — cnooo
ro M ru
SU- LL
(- 1-
E
ru ru j- —
I O O O < U. ^
fU —
i
M —
1
ro
~
0 ?
£ 1
— cn CD r- (D M
— o o o — o
1 1 1 O O O < LL.
O 3-
— 0
O
ujj-ii ru i — ru
z
1 1 i 1 O O O < U.
t~ <->-V _1_I_JZ
<<<< _J_I.JO
CDCDCDCD — — — CO l
DD^DOOOIIIO
BQ !/!<<<< _J_J_lOO
> CD CD CO CD — — — i/l l/l
OQQO_)_J_J(_)U1-JXI — •
3 Q
- Z
Z <
UJQ.CLCLCLCDCDCDtJOCJUll/1
OTUJUJUJU-J—I-JDDDU-U.
-128-
-------
If) O CD —• iT) Ol if) 1C (71
o
in
Q
in
cnOJO) — r- en fo — o
ru — ru — rorururu3-
r-
cn
00
ID
i
S
o
LJ
Q.
a:
LJ
r
u
S
u
^
kNOGRAMS
Z
••
u
_J
*^
cn
u.
o
>-
Q
D
1—
3MPAR I SON
u
cr
z
o
CD
in
3-
ro
ru
-
0
en
00
t-
CD
in
3-
ro
ru
—
RESULTS PERIOD
Q
CD
CD
en
(XI
i
ro
in
3-
in
S
en
ru
ru
3*
03
CO
ru
1
ro
ru
in
171
ru
3-
ru
i
en
o
i:
m
Nl
Q
a.
LJ
Q
r z
_ ro
in in
ro ru
ro in
O CD
0 0
ru ru
r- —
ro ro
in ru
— (XI
co in
r- ru
en —
1 (XI
o ru
(Xl —
CO CO
ru ru
3- CO
(XI —
3- en
en ro
(XI 3-
CO 3-
o en
3- ID
— ru
E 0,
i i
U- b_
H- h~
>- >•
CD CO
NI M
Q Q
D. CL
LJ LJ
H Z
ce o
(XI —
r> 0
ro r>
ro 3-
ru r-
en oo
— ru
ro ru
3- CD
00 0
ro o
o CD
ru 3-
ro in
ru in
co in
— ro
ru i
3- 3-
ru ro
00 ru
ro CD
en o
ru ro
8?
O 1
ro i
0
ID 3-
— ru
en >-
l O
en <_>
o —
H Q
D O
Nl O
Q _l
< _J
Q_ 03
LJ _l
cr o
_i en
ru ro
in —
ro o
ru in
ro en
in co
ru ru
oo en
ro r-
in ro
3- CD
— 3"
— 3"
en en
ro
r- CD
ru
i i
ru —
CD ro
— eo
03 —
O 1
0
3- ru
r- in
o ro
3- LO
— ru
h- t—
O O
o o
Q Q
O O
O O
_J _1
_l _J
CD CD
_l _!
_j in
?
1
1
t
ID
UD
r;
r-
in
L:
3-
cn
in
r-
00
r-
00
G UCD CAHILL SMDAY
1
i
o
CO
1
o
ru
i
I
I
1
in
ro
1
r-
ro
$
CD
1
U-
in
i
u
Q
O
D
o
CD
.n
CO
ro
CD
ro
ru
en
3-
in
ro
ru
in
ru
ru
r-
ru
ru
en
ro
in
ru
en
r-
en
o
en
2 ME AN (WITHOUT I )
in
(V
-
3-
3-
r-
ro
in
to
ro
r-
ru
in
ru
en
to
ro
CD
o
ru
r-
STANDARD DEVIATION
ro
CO
CO
(XI
S
ru
ro
3-
00
ro
CO
CD
cn
00
LO
ru
o
ru
OD
r-
eo
en
o
in
0
_j
T.
0
Z
<
£
X
0
ru
in
ro
cn
0
ru
in
00
to
cn
o
LO
in
ro
ru
00
ru
r~
STANDARD DEVIATION
0
LO
in
3"
ro
ru
-
o
cn
CO
r>
(D
in
3-
ro
ru
—
RATIO TO Z PERIOD
o
Q
3-
co
o
i
CO
o
ru
0
*
3"
O
*
LO
0
ru
3"
ru
—
'
ro
cn
*
3-
cn
o
ro
o
ru
<
i
0
>•
CD
=3
Nl
0.
U
Q
O O
o o
r z
So
[-
o o
r- to
co ro
o —
CO CO
0 O
ID fO
r- to
0 O
ID —
r- CD
O 0
ru ru
CD ru
0 0
03 00
CD 3-
ro ro
j- in
o o
i 01
0
cn in
0 0
ru c-
r- ID
o o
r~ ro
cn r~
*
3- 0
o cn
— o
*
in m
o o
§!S
0 0
*
r~ ro
0 —
1L 0-
i i
h- 1—
V V
CD 03
3 D
Nl Nl
Q. 0.
LJ LJ
JL Z
O O
in o
cn in
o —
tr u
in 3-
cn ro
o —
r> ro
CD (XI
o —
5 ru
0 —
CD r~
co ru
o —
— ru
cn ro
o —
r- cn
co cn
o —
ru o
cn ru
oSS
o —
O 1
o if)
.
03 —
£7) —
o nj
*
CD ru
in r-
CD —
o —
CO 1
*
* '
o
*
03 ru
0 —
in i-
1 0
Q —
>•
03
D 0
Nl O
Q- OD
Lo _J
tr o
o o
_i in
01 O
LO in
o —
o —
ID ro
o —
to ru
o ru
*
00 —
r> CD
o —
LO —
cn ru
0 —
00 0
3- cn
o —
co co
ro in
3- 3-
ro ro
o —
i i
in CD
0 —
co ru
J- (C
0 —
ro r-
^^
ID 1
Q
*
cn nj
o OD
*
r- ro
o —
t— ^—
0 0
Q Q
0 O
O O
03 m
_i _i
_i in
0 0
CD ro
o —
i i
i i
i CO
o
i i
1 O
CD
O
cn i
CO
o
5 '
— i
CD
O
r- i
0
ru in
0 —
3- 1
f
O
r- o
in CD
O 0
*
CO I
o
ro in
0 0
3-
0
>•
<
r b.
in in
_i _i
X I
< <
O <->
D D
O —
0
0
0
0
o
0
o
0
0
o
o
o
0
o
o
o
o
o
o
o
0
—
o
0
o
o
o
0
o
o
o
0
o
MEAN! WITHOUT I) 1
cn
ro
o
3-
ro
o
0
ro
o
cn
o
3~
ro
0
ru
o
3"
CD
O
ru
to
CD
0
r-
o
in
in
o
ro
CD
0
in
ro
o
o
in
0
03
0
CO
ro
STANDARD DEVIATION 0
L.3- — ineoior-iocD
L_cncncncncncncncn
UOOOOOOOOO
0
• ruroinr-r-rO3-co
> — o\r>(Drf>~-OJUi
LJ3-ru — ruro3-r-ru
• 03-0100 — coior-o
orororururorururu
oroooooooo
UOOOOOOOOO
LJincoiDCD — rom —
DLrocnoioincncor~
en o o o o — — — oo
(Er-incn — 3-co — o
ooincoo33- — 3-ru
rrtnr-cneoorjointo
t— ruoODinrucno —
o-coinooo3or>3-in
Ljrocooror- — ru r-
oroicnin — roj-ruruo
LJ II
REGRESSION VS Z INI
D EPA DZUBAY MDS-A
M EPA DZUBAY TF-M
N EPA DZUBAY IF-P
R EPA DZUBAY MDS-S
C LBL LOO DICOT
L LBL LOO DICOT
S LBL LOO DICOT
G UCD CAHILL SMDAY
I UCD CAHILL SFU
-129-
-------
r-
01
CD
i
CD
cn
o
o
LJ
Q.
ct
LJ
LJ
r
o
§
o
5
C£
O
O
z
z
=>
o
LJ
1
_J
U_
O
1
cn
oc
Q.
o
CC
h-
z
Q
O
vD
in
:r
ro
ru
-
o
cn
00
P.
to
in
3-
ro
ru
—
RESULTS PERIOD
a 3: z or
cn 3- — ro
ro — — ru
cn i in ro
ro ro —
i m o ro
— ru ru
CO o co ro
in ru ru —
o j- co ro
ro — — —
i o ru CD
i cn — o
CD ru o a
I 1 — CD
— CD r- ru
to
t r^ CD in
CO 00 I CD
ru ro —
j- o ro co
ru in ro ru
I CO CD C71
ro oo OD r>
in co o r>
i i i i
9 *~ *~ P
>->->->-
OD OD CO CD
Z> D 0 D
ISI M M N
a, a. a. 0-
o _j cn
o en m
ru ru
ro r- in
co ru ro
ru
J- OD I
ro o i
ru —
in r- i
ro in ro
co ro —
i t t
— in in
o r~ o
ru ru
ru r~ co
£5 '
i CD ro
p *
i OD in
1 CO O
ru
o o o
o o a
o o o
o o o
CD OD CO
O
!
,
1
t
ru
CD
i
co
i
r-
r-
1
cn
i
oro
cn — r- co cn J- — r-ro
— — oooo — o —
orzrro-Jino —
3- o in 3- co in j- i i
co r- in — cn ^ ru
— o o — o o —
* *
ru i ru j- j- J- J- i i
cn r~ co co ro r-
— —oooo
i — co 3- i- in 3- i oo
oo o ru co to ru CD
0 — — O O — —
r~ in rji in o j- i i i
3- oo — tn co ro
ru o — o o o
o cn — 3- o r~ i i i
r- r~ cn r~ ro in
— O 0 0 — 0
*
t O O O O O 1 O 1
o ru to en r- ru
_ o • o —
* *
t CD ru OD ro co ro — i
o — — in in in r~
00 ro co ru — CD i i
oo co CD CD ru r-
ro o o o o o o
*
i i ru co i i i i t
o —
*
fom — cnruroooco —
oo — r-r~rocoro —
ru — — ooooo —
* * * *
i o ru CD ~ en — i 3-
ru 3- o i- J- J- ro
— 0 — — 0 — —
* * •
r-ro i — — CD — t r-
J- CO OD CD ro OD 01
— — oooo o
*
— o ru in — — i i CD
co r- — co co CD in
o — — o o o —
* *
i in — r- i — in ro i
o o — o — o
* * * *
r~ — — CD i o CD in i
— 01 ui oo j- r- ro
ru o o o o o o
t *
CD j- j- i in ru i i
ro CD CD in co
— — o o o —
-
l I 1 I O O O <
lflLi.Lj_tr>OOOOD
Qh-l-O — EU.
r i: Q a a cn cn
00 CD CO CO — •—
DDDDOOOIX
MNISINOOO<<
OQQQ-J-I-JOO
O-CLO-O-ODODODOO
LJLJLULJ_I_I_IDZ>
0 CD
0 in
— O
O —
o in
— o
o in
0 CO
— o
0 f
o ru
— o
0 00
o r>
— o
o ai
0 3-
— 0
0 OD
o in
— 0
O CD
o tn
o —
o ro
O CD
— o
o ru
o in
— o
o r
o 3-
— o
o cn
o in
— 0
O OD
o ro
— o
0 3-
— o
0 0
O CO
— o
o —
o r-
ME AN (WITHOUT G, I) 1
STANDARD DEVIATION 0
Luinooinina- —
L_rotDLOcnroruco
LJ — r-aocoj-r~ro
oooooooooo
tr
o
• in o m co QD tn CD
>oinoocDoroj-
cjruooro — tDtnro
•j-r^cDj-j-njj-oo
tn —
a:
orutDcncQj" — cn
ccoiroru — ru — ru
UIOOOOOOOOO
Ljcococn — oiaco
o_tncDODCor>rocD
oroj-cDcrroj-ro
cno — — oooooo
cr — inrorucoj-co
ocn — — ODcoroco
orrurooruj-or~
CLj-3-cDor-r-o
Ljcncococoj-inru
cccninrorocno — oo
LJ ru i — —
REGRESSION VS Z INI
D EPA DZUBAY MDS-A
M EPA DZUBAY TF-M
N EPA DZUBAY TF-P
R EPA DZUBAY MDS-S
C LBL LOO DICOT
L LBL LOO DICOT
S LBL LOO DICOT
G UCD CAHILL SMDAY
I UCD CAHILL SFU
-130-
-------
p-
1
00
<
E
(D*
in
8
u
CC
UJ
t-
r
u
m
J
O
in
E
1
i
z
z
2
o
b
s
m
§
1/1
cr
2
O
£
,_
z
-
o
L/>
5
10
ft
J
2
ru
-
o
cn
GO
p-
10
J-
M
ru
—
8
£
in '
in
UJ
(r.
OEZCCO-JUIOX — T*
o r- ID — m fo co i oo tu CD nj
co co cn p- cn cn co P-P-IDP-
o i j- in cn ci o i — ru — co
co op-r-r~(D p~ in 00 CD
— — CD ro ro ro o ru cn —
QD to in J- co — i i ro ru ru M
o — o p- o — oo cn — ru
§j~ o 0*1 in OD i i r- fo in 00
o — co ru — o o cn cn
i j- in — mo i j- o ro CD ro
p-iop-cno p- to cn o o
i — o ro P- — ro o to cn j- m
cop-ioiDr-p-ioinj-ioio
— in— loruop-p-oruruo
coiotDioinininj-j-iOQDCD
' 'MOD' ' ' o CQ o 5 -
ru _, _. _
ID cn ro m — j- to — ODP- — P*
ODODioiDODODcntoior-cnru
"~
i p'focnroruru.ytDruo'ir1-
oj- cnruruMcncncnruj-
j~ j" cn ro ru CTI i co — ru o j
ru ru ru — ru ru ru ru ru ru ro
i(OOOj-iinp~o)oini i
ruoo M co cn — cn
003-fUP-iCnj-iD — 3-1 I
cn cn cn j- ro j- ro j- in
runjruru ro ro ru ru ru
o ru in to i — p* t on ro i i
o co cn r- oo to 10
- ru j-
I 1 1 I O O O < U.
inu.u.u")oooo_j'D
Q t— t— Q ~ — — E t— u. in in
E EOOOint-cn_J_,
>->->->- _J_I_)ZZ
<<<< _I_1_IOO
Ullsssillll
QQQO_i_)_JOOOII
'
fLCLO-CLCDCOCDoooinin
LJLJUJUJ-l-J-JDDUU.L..
arzoro-iLnoi — -> x
ru
ru
ID
M
00
cn
r-
-
ro
o
j-
o
S
s
ru
ID
in
—
ru
00
o
^
ru
£
ru
ro
ru
p-
co
_
Ul
or
UJ
s|
in
_j
_j
— •
z
<
£
M
O
2
~
in
ru
;:
m
£
in
m
CD
p-
ru
s
o
ro
?
ID
Z
O
<
>
1
z
<
in
Q
o
£" 5
m to in
r- — —
ru — —
cn — ^
o — —
o — —
Sir
in ru o
CD — —
j- j- cn
CD —
(D tO 00
O
£ffl "•
to r- ID
o —
- ru
ru
— OJ
£0 ru
J-
ru
P- CD —
CO —
_§ i
5 ^ or
• < u
™ -. £L
- a
_j
E CC M
Q O O
" ? ^~
1 * 0
u,, -
x or
ro o o cn
O O O O
O O O 00
o r z a:
Sin in in
0 — CO
— i ru in
o ro cn
r " ?
o a r-
— — o
in o ru ru
o — o r-
_ _ _ o
£§££
o — — o
* *
1 SF:*
o o o
i j- r- co
ru o cn
— — o
— in cn r-
ro o cn o
— — o —
1 • gp
ru o
j- co r- cn
0 0 P- P-
— — o o
* *
' O ID fO
— 0 CO
cn ru ' CD
O — O
oo ro cn o
oo — cn oo
o — o o
*
0 CO CO
— o o
O o O CD
— — — o
• *
in in o CD
— cn — CD
— o — o
i i i i
a t ii S
1111
O D Q Q
CL CL CL O.
o E z or
— 0 ID P-
o o o a
0 _| U10
j ru 10 i
— — 0
00—1
o o o
777
_. - -
in oo i i
o o
— -
ru —
- -
r- CD i oo
— — 00
— — o
gg~8|
— — — o
S— nj to
00 Q> P*
o o o o
i i i 00
CD
0
CD ru r- j-
Ol 0 — P-
o — — o
* •
tfi j- r co
ru r-
— O — CD
— ^- — O
ID — i 00
— — o
— in CD
— — o
ru ru co
— — o
— ru
_ _
o o o <
— — — E
O O O I
_J _1 _J O
CD CD CD U
o _» m o
o ru
o c
J Ol
i -
*0
* •
cn to
o o
m o
o —
— o
CO CO
o o
ro co
o cn
— o
• •
— o
I
o —
Sic
o o
in o
co o
0 —
ro CO
p- 00
0 0
ro j-
CD cn
0 0
*
— CD
CD CD
O O
? ?
O O
881
O O
* *
— cn
cn r-
O O
CD cn
O O
CO P-
o o
u.
_i _j
i i
L) O
§§
i _
CD CD
0 0
-J *
88
O —
7 7
eo 8>
O 0
co cn
o —
7 "
oS$
o o
* •
in ru
ru ru
— :
S8
o o
BfS
— —
P8
O 0
- S
— —
n o
cn —
o —
o r^
7 "
oo ru
o —
1 '
1 '
1 i
ru j-
in in
HUDSON
HUDSON
tn tn
-> *
O Oi
o —
o ru
0 —
o r-
o —
— o
o —
— 0
o in
o —
— o
o —
o —
— o
o —
o ru
— o
o ro
o —
— o
o ru
o ru
— o
0 CD
o j-
— o
o ru
o ru
— o
O (D
o —
— o
o —
— o
o m
o —
— o
o m
o ru
— o
o in
o —
— o
o cn
o —
— o
-|
LJ <
8=£
3°
< O
- Z
z <
< t-
LJ in
fc £ jjj | £ 5 S jjj S £ 5 S S
o • ...
ooooooooooooo
g
u
ucDcoiDj- — mo — — f — m
o •
•j-cnj-cooor-io — o — j-r-
E ~
cr
§S;r?foinf'i5inl:??-o> —
OO — OOOOOOOO —
DC ......
UJOOOOOOOOOOOO
ocnoocorururucncncnionj
_j . . . . . . .
o^— S!ninop^Scn""cotf1
cr j-p-rjcocorururuinro — to
cc - •
ujcntncnj-mj-op-iniDoru
afflfninr-SrucncninSp^ —
unj^cocnoficnrotD — ooror^
orcoruco — ruoor-aorur-ro
ui i — — — i i i ru —
z
ruj-
I I I 1 O O O < U.
> co CD cn CD — — — tn m
DDDOOOOIIIQD
ZMMNiisjOOO<<
-------
i-
en
00
ID
i
0
o
cr
cr
H-
Ul
>• V V _) _l
CD OD CD CD — —
PvlP>jPsJNOOO<<
OOOO_1_J_JOCJ
0.0.0.0.00000000
LJUILJLjJ_J_J_JDZ>
ojezrro-icoo —
"
CO
o 3-
in
00 00
PO
r- ru
PO OD
in
— co
r~
» °°
J- 00
ru in
PO
— en
on r~
PO
S"1
in co
co
— ru
en ru
o in
PO r~
j- ru
ru
in en
in
Z ME AN (ALL SAMPLERS)
STANDARD DEVIATION
— PO
in
CO CO
PO
CO —
r- —
in o
in —
— r-
r- 00
3-
co en
3- CO
PO
3- ID
PO in
o r-
co
co
CD ru
en ru
CD CO
PO ru
ru
in
X MEAN(D,M,L,G, I )
STANDARD DEVIATION
o
CO
O
co
in
ru
0
en
00
r-
co
in
3-
PO
ru
—
RATIO TO 2 PERIOD
poruoocoocoinru
ooooooooo
en — oooocno — OD
0 — — O — 0 — — 0
ozzrro_jLOo —
cnr-r~cor-cnru i in
oooooocncn rji
— — — o — o o o
cnpocorucncnru t —
corupoocnoocn r-
o — — — ooo o
* * *
i — o 3- 3- r~ o i in
en o r- — — — on
o — o — — — o
* * #
co in co 3- r~ in i i p-
— — en CD o o r-
— — o o — — o
* • *
oo3-por~coin i PO
cnocncnocno —
0 — 00 — O— —
oocompocoin t ruin
ooruoooDonoD ru —
o — • o o o o — —
* *
onco — oo — cnincnpo
CDPor~cnooDcnpooD
o — oo — oo — o
* * *
coocooncopopo — co
ruruonoioioooo — oo
— — o o o o o — o
• * * *
— i 3- PO i i i oo a-
o — o o o
3-incncocnr~0)3-r--
— rucoooenoicn — en
— — ooooo — o
i pO3-cnmcn3-oco
— en on o co en — co
— oo — oo — o
j-j-rucnor-inina-
en — oooocncn — en
0 — — 0 — 00 — 0
OIOPOCOOCO i in —
OD — coco — o — en
0 — 00 — — — 0
* * * *
PO in — oo i opoinco
0 — 00 _ — — 0
* * *
oooru— i 3- r~ — co
oioocn — — en co
0 — — 0 — — 00
* * * *
j-r~inocnpO3-inru
o — focncncnoDcnoD
— — — oooooo
* » • »
"fTTfoooS
OK- h- Q — — — ZL»_
r £ o o o en en
>->->-> _j _i
CD 00 00 CD — —
DD3DOOOII
OOOO_I_I_IOO
Q.O-CLQ.C0630DOO
o r~
o —
— 0
0 OD
o o
— o
o —
o ru
— 0
o in
o —
— o
o in
o —
— o
o OD
o o
— o
0 CD
O —
— O
O PO
o ru
— 0
o co
o —
— o
O PO
— o
o r~
o —
— o
o en
0 0
— o
o en
o o
— 0
o —
o —
— o
o en
— 0
0 —
o —
— 0
o r~
o —
ME AN (ALL SAMPLERS) 1
STANDARD DEVIATION 0
u.
u.
o
o
cr
o
s
CO
E
cr
o
cr
IT
CJ
Ul
O.
OL
O
gj
Ul
Q.
Ul
REGRESSION VS Z INI
ru in oo
en en ru
000
3- r- f\i
r- r- PO
r- ru in
r- co 3-
ru
ro fu o
o o —
ooo
ID tD if)
nj o> —
— ru 3-
cn in —
po co PO
po ru o
— co in
oo in CD
co o m
ru r~ ru
< *: a.
t i t
CO u. U
O I- t-
>->->•
CD CD 00
psi M rsi
OOO
Q. Q. Q.
LJ Ul Ul
o r z
c. on oo
on on en
ooo
oj in r-
o on PO
CO 3- I-
r~ ru 3-
PO — OJ
ooo
ooo
ru co en
co ru in
co r~ —
en oo o
— ru o
PO — ru
r- oo on
3- oo —
GO 03 OD
o 3- en
i i i
en t- i-
l 0 O
en o u
Q
zoo
CD
ooo
Nl O O
O _l _l
O. CD OD
Ul _l -J
cr o _i
en CD on
ooo
in 3- oo
co in in
en po r~
in co PO
ooo
ooo
PO CD ru
PO 3- r~
o PO oo
in co oo
in co —
3- co PO
r~ PO r~
0 — —
in PO j-
— CD —
i- >-
0 <
0 O D
— z: u
o en co
Oil
o < <
_1 O U
_l O O
OD O U
_) 13 D
CO 0 —
-132-
-------
o
in
njnjrunj — — — — ru
Q ru^-iocoro — rocoru
in rururu — rururuoru
r-
cn
CD
V
z
ID
[R; PERIODS I
LU
h-
O
CD
O
I/)
n
i
o
z
z
z
Nl
LJ
O
cc
^
fc
>-
s
1—
CO
g
in
Q.
z
(J
cc
LJ
h-
z
-
o
^
10
m
3-
ro
ru
o
cn
CO
r-
10
in
3-
ro
ru
—
fc—
a:
LJ
0.
in
t—
_j
in
LJ
CC
Q
ID
ro
ID
3-
10
8
i
o
1
—
J-
t
r^.
^
*.
i
cn
in
ro
3-
<
in
Q
Z
>
CO
D
N
Q
Q.
LJ
Q
Z Z
J- ro
tO J
i ru
in
to ru
j- 3-
ru 3-
in in
to to
to 3-
cn co
— ru
in ID
to r
ID O
ru ro
i —
ID
o ID
j- to
in in
ID 1
ID
ID 10
ru to
^ £•
O 3-
m j-
CO -
— ru
JC Q.
U. U.
»— t—
V V
< <
CO CO
M N
0 Q
ru
— ro
ro 3-
to 10
r- t
ro ID
CD —
ru r-
3- i
ro
fi '
r~ i
ru
ir
i i
in i_
D —
Z Q
>-
CO
D O
Nl O
O _1
< _l
CL CD
LJ _l
CL 0
_i in
3- ru
3- 3-
in in
j- 3
r- ru
3- in
in i
in
CO I
3-
ro i
ID
o —
f 3"
ru o
ru ro
i i
ID r-
3- in
j- r~
ID ID
— ru
ID r-
ID I
ID
in 3-
in ro
— Ol
10 in
o —
in ID
K- 1-
O 0
O D
O O
O O
ajoj
-J in
o —
i —
ro
i in
ru
1 0
in
in
ru
— ro
ru f
ro —
— ru
— 3-
— ro
ru 3-
ID cn
— ro
o ru
ru 3-
ro cn
ru ru
co r-
— 3-
in ID
— ru
10 in
— 3-
l CO
>
O ID
Z L.
in in
_i |
i i
u o
Q Q
U O
0 —
o
CO
ro
3-
3-
3-
CO
3-
m
ro
CO
ro
ru
ro
CO
ru
ID
cn
ro
0
in
0
in
0
in
ro
in
3-
3-
ro
in
cc
LJ
T~
in
z
NJ
ID
cn
CD
cn
ru
-
cn
r~
„
—
CD
Ol
cn
ro
£
r-
z
o
5
u
D
Q
CC
Q
Z
in
ID 10
ro
cn ru
ro —
CO ru
3-
ru —
ro —
r- —
ro ru
r~ ru
ru -
i
ru —
CO CO
f\i
10 ru
ro —
3- cn
in cn
CO O)
3" —
in co
ro — •
ID CD
3- —
ru r-
ro —
•z.
0
l~
Z —
o uj
• a
£ CC
Q a
z <
< 1-
LJ in
z:
X
o
^
ID
in
3-
ro
ru
0
Ol
CO
p~
10
in
3-
ro
ru
—
D
O
CL
LJ
0-
Nl
O
o
CL
O
r-
0
Q
*
O
•
S
o
*
nj
r-
o
in
j-
„
'
ID
O
*
1
3"
cn
O
CO
CO
o
1
ru
ro
__'
ru
_«
<
in
Q
Z
>-
CO
D
Nl
Q
Q-
LJ
O
O O
o ru
0 —
z z
".-
o —
*
I ID
ru
in ID
o cn
— o
CD ru
o —
in ro
cn ru
o —
o j-
in r-
O 0
co ru
o 3-
3- cn
cn o
o —
* •
1 S
— *
ro ro
o a>
— o
* *
aj to
0 —
— —
*
ru i
ro
"""
CD ru
ru ru
_
CO —
•— • ru
™ —
— CO
— O)
— o
* *
ro ru
in ID
o o
z a.
i i
U. Li_
t— ^-
V >-
CO CO
Nl Nl
O Q
Q- 0-
LJ LJ
Z Z
o
CD
cc
in
0
£
o
3~
(D
o
o
ID
0
ID
ID
O
ru
cn
o
ro
cn
*
ro
ru
—
p
.
0
o
CO
o
ru
10
0
3-
r-
•
^
S
O
c?
o
*
ID
in
o
cn
r-
0
in
in
Q
^_
m
Nl
Q
0-
LJ
CL
0 0
ID ID
— ru
u _i
ru r~
cn cn
O) 0
o —
* *
ID r-
cn o
c —
* *
ID 3-
O —
*
— CO
3- ro
o to
in to
ru 3-
o ru
*
ID O
r- oo
o o
i i
— cn
3- CO
ro ru
_ _
CO ru
ru ru
^- *~"
ru ru
3- ro
— —
i in
if)
—*
i 10
ro
_!
i r-
3-
_«
1— h-
o o
u o
Q Q
O O
O O
_l _J
_l -J
CO CO
o _/
0 0
CO J-
ru 3-
in o
ru i
cn i
o
*
cn i
, i
i in
in
0
1- 0
ru 3-
cn o
o j-
— o
*
i —
(D
O
t*** —
J- J-
— 0
s?
— o
J- ID
J- J-
•— 1 ^
I 10
ro
0
ID ru
cn 3"
o o
*
ru ID
ro ro
— o
CO I
r-
H
i- >-
o <
O D
— Z
Q in
i
O I
o <
-J 0
-J Q
CO O
in o
0
r-
co
ro
CO
0
o
ID
o
3°
*
ID
o
;
o
to
*
in
ID
ro
OJ
„
•
cn
O
*
,-,
o
—
CD
O
CD
ID
O
J-
cn
0
ro
.
o
o
o
_
*
ro
in
o
ID
L
Ul
•
i
u
Q
O
—
O
O
o
o
o
0
0
0
o
o
o
o
0
o
0
0
o
o
_
0
°
.«
0
0
—
o
0
_
0
0
— '
a
o
„
o
.
— •
0
cn
„
0
o
__
in
cc
u
o_
in
z
LJ
ro
ro
^
10
o
0
CO
0
cn
o
3-
to
0
in
3-
O
ro
ro
o
CO
.
0
cn
ru
0
LO
ro
o
r-
ro
•
*—^
CO
ro
o
ID
f)
CD
r-
rn
o
in
o
z
o
t—
LJ
o
0
cr
o
z
h-
LH
Li_r\jj- — mj-LOCnroo
LtJLDClJliJOtlJP~n~(jOtD
oooooooooo
g
^
• — cocnj-conjroQDco
>rvj[^cntoxcnLOLnLn
ujCDror* — ruiDr-njco
•r-cocnj-CDr-GDrooD
tn
r
a:
cc ID in nj in ro o* r* to cn
Oj- — CDtnrocDCDrorx/
Ccrororo — ronjro — ro
LJOOOOOOOOO
ujinMojonjcDocotn
aoaicnajrojro — nj
ototDootDnjinrocn
ctr^CDoroLOfXJinr^r-
o — — — or-LDincno
CErooir-rocDCDLnLnj-
ujj"f\jinLoro-^LOinro
— — — — — — —
»— ru — CD — j-rootou")
CL ID in oj j~ ro j* — • QO cn
Luonjcrij'mr^ intooj
crcor-ocDCDoaD^-nj
uj — • nj i\i — > i t
t— i i
-z.
-
1 I 1 1 O O O <
Qt— i— Q — — — ^L«_
n z: Q o Q (/) cn
N
> DQ CD CD CD — — •
ZNJNNNJOOO<<
OOQC1Q_J-1_JOO
cn
LJCLCLCLCLCDCDCOUCJ
CCLJUUJLJ_I_J_JDD
O
U
ctaz:za:o-Jtno~
-133-
-------
INTERCOMPARISON STUDY OF TOTAL SE ; NANOGRAMS/CUBIC METER; PERIODS 1-16, MAY 11-18 1977.
SULTS PERIOD 1 2 3 4 5 6 7 8 9 1 0 I 1 12 1 3 14 1 5 1 6 AVG SD
Q£zccu_iinoi —
10 cn p- tn m in in t r~ i
i r- ru nj o nj o i — i
— M P- CD — — 1 1 LO f)
ru ru — — nj ru — ru
p- — m co m fo i IJ-P-
— ru — — — — ru
LQiD.rp-r-r- i in j- r-
ruroruruturu rururu
co — j-or-QDP-ioco i
ru (D i ro AJ to ro i i i
P- I 10 CO I 1 I I 1 1
CDOlOP-J-J-in I ID 1
i CD TO r- ru j- ro i i i
(O ro j 03 in in j- i i i
co — j- r- P- 10 i i i i
, _ __ nj
cn ru co cn i co co i r- t
ru
o co in o i P- p- i p- i
-
1 I I i O O O < U,
a »- »- a -- — — r i- u.
z x: a a a in t- in
V >- > >• _J _) _J
ODIDrDOOOrrTI
aOOO_l_l_lcj(Jcj
<«*<<_i_l_jaOa
CLQ,0-0_COCOmoOCJ
LJLJLJLJ_I_I_JD3D
CJ —
8 5
o
o o ro
— > o
<
to — P- ru ID co
— cn
o ru o ru in ru
— — — ru
ru ru in j* r i
ru
j
ro
0
o
in
r-
f
cn
O
ru
OKI ru — ro P- co
ru ru — o —
o —
CD ru cn ru o ru o
— o j-
•
ro — j- ru cn ro o
to cn
o — •
•
p- — r- o co 01
o
10 ru P- ro p- co o
ru U)
in ru ID ro LO i co
r-
00
0
r-ru co ro j- ru
o m ro co ru 01
o
CO ru CO ru — o
ro
§ § § ?
— I— »- CC O
• < < LJ r
0- - a.
t- LJ LJ >•
§O-O <
-J CD
3 o a o o
z < z < *~ <
< h- < t- O Q.
LJ in LJ in — LJ
g
s
p-
ru
j-
o
i
1
Q
Q-
LJ
ru ru nj
— ro cn
ru ru —
0 0 0
ro uH ro
CO 0 CD
j- ru ru
— CD CD
— ru —
— ru cn
_ _ o
o o ro
— — o
p- ru P-
ao cn o
00 —
cu j ru
o m o
* •
o o o
— p- cn
in ru CD
0 — 0
I if) PO
cn ID
0 0
•
ID J i
CO —
O —
* *
ID ru J-
0) - ID
o — o
P- LO J-
ID *f) 3-
o — o
— p-
— o
ID CD CO
m cn oi
o o o
* •
Ol Ol 1
o o
*
If) O 1
ID ro
o —
I 1 O
< <
m CD
O O _J
a. a, CD
LJ LJ -J
ru ro ru ru
o o o o
Q J- CO O
cn oo t- cn
ffiffi ' £
01 cn oo
o o o
o ro i ru
— o o
* *
p- i i ru
0 CO
~ O
S ' ' »
o o
o o o
ru 01 ID ru
o CO P- o
in in i i
01 cn
O O
* •
J- O I ID
(D CD 0*)
o o o
Sp- 1 1
(D
o o
p- m
0 0
» ' ' '
o
O O ID CO
— — 0 0
cn cn i cn
o o o
01 Ol 01
o o o
'-") O < U.
o a in »-
_j _i
_! _J
_J -1 (J O
53§§
ru
cn
o
f
CO
CO
a
o
'
1C
is
I
<
a
CJ
o cn
o ru
— o
o ru
— o
O LO
o —
— o
o o
o ru
— o
o ro
o —
— o
o r-
o ru
— o
o in
o —
o p-
o ru
— o
0 P-
o j-
— o
o j-
o —
o ro
o ro
— 0
o j-
o in
— o
0 CD
o y
— o
o ru
o ro
— o
o ru
— 0
O Oj
o in
— o
o ro
o ru
— O
z
o
MEANCU1THOUT G, I
STANDARD DEVJAT
U.
8
(J
or
o
u
LJ
O
or
o
LJ
LJ
_J
in
g
INTERCEPT ERF
GRESSION VS Z
CO 10 0 —
p- ro if) ro
cn 01 cn 01
o o o o
cn m j* r~
ro cn cn —
— ru — ru
— tn P- —
ID ru CD 01
o — o o
o a o o
01 nj ru o
CD o — r-
cn ru cn co
o — • o o
CD — r- co
ro — CD r-
a — _ —
m ro p- j-
o ru o —
< r o. in
M rst fsj M
a o o o
o. a. o_ o_
LJ LJ LJ LJ
ro o cn
cn 01 in
cn cn cn
O O O O
ID oo nj
j- cn cn
cn 01 cn
O O O O
_ ru —
3- r o
00 —
o o o o
O) — CD
cn ID nj
O O O
_ — _ o
J 0 CO
o — in
o o o o
§o cn
in 01
ro ro
— — — o
o o o <
O Q O l/l
_J
O O O I
o o o <
-J _J -J CJ
^ CO CD Q
-J _) _J D
ru
%
o o
I
— o
J-
in
o
0 0
o
CO
o o
p-
o o
o
ID
o o
£2
-J -J
I I
< <
l_) CJ
Q Q
- £
a:czcco_imoi~ N
oEZcco-jtnoi-
ccozrzoriJ_jcnoi —
-134-
-------
-
rururoru — — — ru^*)
Q
in
r-
01
ID
— ru
in
8 -
tr
LJ
0.
00
z>
in
r
cc
o
o
z
<
z
<
r
in
-z.
o
in
a:
<
0.
o
o
oc
arzccu-itno —
in oo in in j- j~ .r i i
o CH CD O GO CD QD
in CD nj cn o cn i i
Ol ru 10 10 Cn CO o I f*l
— 01 — — — — ru ru
in oo ru 10 .j- ru in i r-
_ — — — _ _ ai
loio — j-inj-inmr-
rurorururururururu
to — ru ru M ro i i
r~ i to r- i i i i i
to GO m in ro M j- i
r- M ID ai M ru
o — 10 J ^ ro I I
r- — ru ID ID in i i i
rotnooro i rurucnru
— — — — — ru
— coooinoor-ioioj-i
Q
O
ft
LJ
Q.
_
I 1 1 1 O O O «1
inu_U_LOOUOC:j
ai-t-Q — — — zu.
r: n a a a in m
_ _
CO 00 CD m — —
DDDZ5OOOII
NIM!NNOOO<<
OQQO_1_J_J(_)O
___
Q-Q-CLQ-CDCDCDCJCJ
UJLJLJUJ_I_1_IDD
ro
CD
in —
CO —
— M
cn ru
m J-
10 in
ru
r~- ru
ro ru
r.-
in ru
j- ru
in rn
10 ro
ru ru
Co ru
r- —
MEANfWITHOUT G, i )
STANDARD DEVIATION
10 ru
cn —
m jr
o ru
ru
in ro
cn 10
ru
00 -
ro ro
,0
CO ro
in ro
10 m
oo ro
m ru
oo ru
r- —
MEAN(D.M.L)
STANDARD DEVIATION
Q
in
o
LO
in
j-
ro
ru
o
cn
00
r-
(D
in
*
rn
ru
—
Q
0
LJ
CL
IN
O
1—
o
— r-
0 0
f ro
O LO
O O
o 10
*
in ro
— o
i ro
j-
ru co
o —
•
f r~
— ro
— cn
o ro
J 01
— ru
cn M
m ro
o ru
in i
*
— in
ro r-
00
10 ru
o —
»
f 00
— r-
*
r- ro
*
ro r-
cn y
o •—
*
[: !:
EPA DZUBAY MDS-A 1
EPA DZUBAY TF-M 1
m J-
o o
cn in
10 0
0 —
0 O
0 0
* *
ru in
cn —
o —
ID J-
r- —
10 ID
00 00
O O
5?
o o
*
CD cn
o o
ro cn
jr ru
cn CD
ro r-
0 0
*
o in
0 —
* *
uo cn
10 o
* *
oo r-
r- in
0 —
*
— r~
ru ru
nj r-
ro cn
o o
*
tD r-
o —
*
r* r-
CD O
O —
*
P £
EPA DZUBAY TF-P "0
EPA DZUBAY MDS-S 1
10 ru oo
o o o
oo cn cn
o o o
000
CD 00 00
o o o
ru ru ru
cn cn cn
000
10 in LD
co cn oo
000
ru ru r-
000
* * *
r~ — f
o cn —
— o —
* *
cn cn 01
0 O O
0 O 10
0 0 CO
* •
oo r- r-
f-
o — —
• * •
i i <
ID CD 00
ID CD 00
O 0 O
ru CD ru
in r- in
000
in in r
00 CO CD
* *
cn CD
o o
*
1 CD 01
o o
* *
i ro ro
cn cn
o o
* *
0 CO CO
— o o
* * *
o o o
o u o
O Q Q
0 O 0
00 OD CO
10 00
0 0
r- in
o —
i i
i i
i i
ru
*
i cn
0
ru
01 0
o —
CD I
CO
1 1
1 1
, ,
1 |
1 1
1 1
J- ~
0 —
1 |
m
o
<
Q D
U b_
-J _J
-J -J
I X
u u
o —
o ro
— o
0 CD
o nj
— o
0 —
o —
— o
o m
O OJ
— o
o —
0 —
— o
O CT1
O OJ
— o
0 CD
O —
— O
0 0
o ro
o r-
O ID
— o
o cn
— 0
o nj
0 J-
— o
o to
o in
— o
0 0
O ID
o r-
o j-
— o
o CD
— o
o ID
o rxj
— o
0 CD
o —
MEANt WITHOUT G, I) I
STANDARD DEVIATION 0
U- —
UJ U'
u o
cc
o
> 0
• o
(fl
r
cc.
o t
a. o
LJ O
Q_ —
in —
o ro
Sr *
Ul O
o_ co
CC 0
LJ
GRESSION VS Z INI
EPA DZUBAY MOS-A
ro — oo
O 0 0
ru 01 ru
r~ r" tn
oo ro o
ru
co in CD
o o o
o o o
— CD —
— o o
r- ro ro
CD CD 01
000
CD 00 J-
1
EPA DZUBAY TF-M
EPA DZUBAY TF-P
EPA DZUBAY MDS-S
j- in
0 0
in —
o o
M ru
0 0
o o
ro r-
— o
00 0
ro ro
0 0
oo cn
0 0
i i
o o
Q Q
O O
O O
_J _J
00 CD
_l _J
01
0 O O
I
— O 0
J-
o
CD O O
in
°
— o o
o
cn
000
J-
— 0 0
1
o <
O cn in
_i _i
Oil
o < <
_! IJ (J
CD O <_)
_l ^ D
ccarzccu-jino —
-135-
-------
C/l
Q
O
CE
LJ
Q.
¥
o
§
u
I
I
I
_J
Q
=>
Z
O
LO
a:
<
Q.
£
O
O
a:
UJ
o
>
<
en
co
>-
<
r:
ID
— 01
a
o
cc
LJ
o_
ru i ro ru — —
ru j- o — ru —
ru — — oj ru ru i t '
ru ro ro ru — — i i
— i o ro ro ru ro
ru — — o — ~
—.00 — 000
— o rxj — — —
j- o ru i ru ru
ru — ro — i — —
— ru o o ru i — —
co
LJ
— a:
-
i i i i o o o <
1/lU.U.LOOCJCjOZ)
oi-i-a — — — j: u.
r r a a o LO LO
X >->->- _J _l
00 OD 03 00 — —
3D3DOOOII
NIN}NINIOOO<<
OOOQ_I_J_JCJCJ
<<<<_I_)_IQQ
Q-Q.Q.Q.COODQJCJCJ
LJLJLJLJ-J_J_j:3I3
— — — o
ru ru ro ru — — — i
1— O — O — — I I — —
— to i ru — — — i
•z.
o
o —
5§
o
I Q
K- a:
3 o
z
o
- Q
Z CC
a a
- z
z <
< t-
LJ LO
X
o
Q
LO
O
CO
in
j.
—
ro
—
ru
„
—
0
—
en
CO
ror-inODor^j-oo
CD oo oo CD r~ 3- in
in ru r~ j- "• — —
CO 0 00 00 f 3- 3-
o — ooooooo
3- r- ru ru en CD in i i
ro — o ru in OD r-
— — _ _ o o o
ozzcccj-jcno —
O O O O O O 1 1
o o o o o o
— — ru o — — —
* * * * •
O 1 O O O O O 1 1
ru co ru co CD CD
— — — o o o
* * # * •
1 O O O O O 1 1
ru j- co ru CD
— ru o o — o
* * * •
O O O O O O 1 1 1
ru CD CD ru ru ru
— 0 0 — — —
o o o o o o i i t
o m in o IP in
— — o o
* •
I ID CD — CD I i i
to to en ro
o — — o —
* • » *
1 O 0 O 001 1
000 0 O
* * * *
O O II
in in
to o o to o o o
II 0 1 1 1 1 1
. °.
o ru
o r~
o r-
— o
0 CO
a in
— o
o en
o 3-
— 0
o ru
0 CO
— o
o —
o to
— o
o in
O f
— o
o 01
o in
— o
oro u_ co cn ru en ro ~ in
OCD u_ — inrotfioocncn
ui co r^ co ro r* 3~ 3~
oooooooooo
0 — CC
or- o
CJ
>rorocooororoo
o— LJj-cn — r^j-coj-
o j" o
•o — — oooooo
— — C/l
— — — — J-
0
o
cr
LJ
0.
O
o
r^ r- in r~ oo oo co
— — r^ — in in in
o o i
in in
_ o — o — —
r~ o
CD o
ro r~ r~ r-
ro 10 co CD
o ru — o o
* * * *
o o o o o
o o o o o
— o ru — — —
* * * *
o o
o o
ru o —
ro r- in nj
co j- cn ro
o o
o
o
-
o
0
0
o
_^
o
0
—
o
o
o
I-
r-
o
s
ro
CO
o
. — .
r-
o
3-
ru
o:
o
o:
ttl
UJ
oe
o
— 1
CO
o:
o
ff
LJ
ID CO
OD ru
— r>
o o
3- 3-
— 3-
3- cn
o ru
— o
ro ru
o —
—
3-
O
O
fu
-
ro
CO
CO
o
00 —
r- ro
ru ru
o o
o cn
0 —
3- cn
o o
CD cn
3- ro
CD O
cn
ro
ru
0
CO
3-
0
in
cn
to
o
r-
CD
0
CD
00
ru
0
00
CD
ru
o
o o
o o
o o
o ro o o
in — ru r~ co ru
O-COj-r^cDrorocn
Ljcnruoooroinin
— — — — o
ru o o ru
LO
a
o_ LO i— »— i— v
i i o o o <
u. co o o cj Q 3
I- Q — — — £ U.
r Q o Q co co
>->->->- -J
00 00 00 OD —
DDDUOOOX
NINININIOOO<
aaoa-j_i_io
<<<<_J_I_1DO
a 0-a.a.cocaoDocj
LJLJLJLJ_l_J_IDr3
o en
o
o —
<- LJ
Z) O
So
t- LT
LJ cn
i:
*->
a:
LJ
t—
~z.
Nl
LO
Z
O
LO
LO
LJ
CC.
0
LJ
a:
0
<
i
LO
a
z
5
m
D
Nl
Q
^
a_
Ul
Q
ru
i
z
i
i ,
^_
5
00
D
Nl
Q
<£
0.
LJ
SI
O
1
0_
1
u.
\—
5
00
D
Nl
Q
-------
< _ j.
&
s
Si
§
ID
in
f
fO
ru
-
o
cn
CD
p"
(D
in
j-
ro
ru
8
DC
£
tn
RESUL
s£
R '
ifi
.,
, j.
, ,
in M
i ,
£jg
1 S
— 00
CD p-
fU
cn r-
r- oo
ru ro
' CD
£5
J- ID
SCO
r-
< z:
ft
>- >-
< <
ss
(SI IS)
o a
2 2
LJ LJ
sHs
ru op in —
JT ru co cn
in co j CD
0 J- 0 00
ID fO CD CO
o ro FO in
a s s s
o — o —
in r- — M
CD y cn j-
° £ S ft
— ID 1*1 ID
in PO in in
o ru i f
° ru
j- o 01 r-
ru 2 ru ro
i p- ro ru
QD m in
ru ru
o — ID ru
— ci s- in
Cf) O (U
r- 01 i 10
ru r- j-
m y m
ru j- i ru
r~ j- cn
o. in i- t-
i i O O
u. tn u u
h- Q — —
r o o
>- >-
< <
3 5 O O
SN O O
O _J -J
a. o. OD oa
LJ LJ _l _)
8 '
8 '
M i
i i
i r*
Ol
P 0
M M
. CO
£ffi
cn in
ru
1 O
in
S5
r- o
O 1
ru
»- >-
o <
o o
Q l/l
J
_J
§1
«t
-J LJ
38
Sro j aj
cn o> 01
o PO M r*
f p- co co
ro — y ro
ID (D r- in
ID M OJ 00
— CD P- —
cn o 01 P-
£o — cn
p- oo j-
O ID O —
ru cn — —
CO — — *
00 M O —
ru ru — ru
r- co o o
— o ru ID
o 01 cn —
in to M ao
j- ID o ru
CD cn m ru
ru ru — ru
O f I I
m o t i
ru j-
b.
_/ z?
»- i/i _j _i
_l _l Z Z
_J _l O O
sils
0 0 I I
O (J l/l 01
D D L. Li-
S
S
CD -
PS
CD I
if, *
o ro
r- M
m o
8?
Kl O
ru 10
in a>
ID —
o m
ru "
j j
J- M
KS
ru m
o m
ru
cn r-
ru
o m
cn in
r —
•HALL SAMPLERS) 160
VNDARD DEVIATION 37
< K
LU LTI
M
J CO
f. «
ru J-
p- ru
,,
p- r-
(D FO
cn ID
jf. *
ru FO
r- in
ro o>
P- ID
ru o
p- ru
* a>
ru
oo j-
3- f
ss
5S
ru
— ID
CO CO
ru
r- tn
(D ID
co in
3- —
S
o
>
ID
in
r
ro
ru
-
o
en
m
p-
l£
in
j-
ro
ru
8
rsi
O
RATIO
j-foj-sDtDnjj-— rifoai —
ruioinp~ooincnLOo — IDLTI
000000000000
r-inmj-r-o — roooor-ro
coruocn*— rufop-cooDCOo1
o — — o — — — ooooo
arzoco-iinoi — ~3*
ro — 01 01 nj — — i CDOOCocn
co — coooj-j-j- iDP-r-p-
o — o o — — — oooo
oo i p-roincQFo i cnffiju
p- rop-oo— p-cnoo
o — o — — — oo — —
1 gSSSJS ' SSPffi
— oo — — — oooo
ro FO j- cn ID — i imcnrop-
p-njcnp-o— cnp-roo
o — oo — — 00 — —
-njcoino-1 iCDin-o
o — cn cn j- j" P-CDQOP-
— — oo — — oooo
* • * •
— o — — — 00 — — 0
o j ; K s £ 8 8 o _ ; :
o — ooooooo — — —
, ,01-1 1 .OOID-ID
— — 00 — 00
cnp-j-j-p-oirup-— inj- —
p-m — cocnorucooor-CD —
o — — oo— • — oooo —
* * *
I njJ'"~J'lD-"COJ"~OCD
PO— • — — — Kl(DP~OlOr-
OGD i rominroruj-j-cno
OIFO cnnjruj-r^p-cotDcn
O Ol — J" J ID I J-ruiOJ-O
o FO — o ru ru mooinoo
imomioai — J-ID« i
— oo ruj-ujp-r-
cn — cocoirururoinrui i
cnj-ocn — ru in P- 01
o — — o — — o o o
oocDoioruinjOOi i
o — OCTI nj r*i mm
— „ „ o — — OO
- njj-
I I I 1 O O O < U.
U")U.lLLOLJO(_>O_J3
>•>>-> _1_J_I2Z
<<<< _)_j_jQO
ZJZJODOOOIIXQO
(NrsirsiMOOO<<<^D
QDQQ_J_J_1O(_)OXI
Q-CLQ-Q-CDCDCDOOOtnLn
LJUJLJLJ-J—'-lDrSDLuU.
o in
o ru
— o
o cn
o ru
— o
o cn
o —
— o
§8
— 0
O O
o ru
— o
o in
o ru
— 0
o CO
o —
— 0
0 0
o ru
a cn
o ru
— o
o j-
— o
O f
o ru
— o
0 ID
o ru
— o
o r-
o ru
0 00
o ru
— o
o cn
o ru
0 ID
o ru
— o
JtALL SAMPLERS) 1 00
WDARD DEVIATION 0.23
< t-
UJ CO
b_ 0
(J o
i
°
> ro
LJ ru
f
X o
LJ 0
LJ O
Q. ID
lo -
CL •
»- ro
CL 0
ae ro
LJ FO
t- i
z
<
S
INI
>
l/l <
Z fsi
O O
tn
LJ a.
g u
£o
,.-.0-0?-,0
00000000000
— rucoruoMooruruj- —
SSCKSSCSRSK
— oo — o — oo — — —
ooooooooooo
sasaSPBSBO
roo — ro — ruj-p-oimiD
ro — ooj-ruroru — cn — in
i£§»i!H5E«!i
ruiroru — j — — ID in
z o. cn t- i- K- >- njj-
I 1 1 O O O < U. • •
SSS ziddSi
S|SgOOl|lQQ
&&&555SSS2SI
-137-
-------
— ru — ru — — —
r-
r-
01
00
i
r
co
i
R; PERIODS 1
LJ
£
O
CD
NOGRAMS
z
-
cc
CO
£
8
t—
JMPAR I SON
£
H-
Z
O
o
CD
in
3"
M
(XI
O
Ol
CO
^
ID
in
3-
M
(XI
—
RESULTS PERIOD
ouzccu-ic/io —
in 3" CO J- ID CD 3" 1 CO
co ro r- r~ in in in 3-
— ru — — ru ru ru —
in o co in — ru r- i in
STANDARD DEVIATION
o
o
to
in
3-
ru
0
cn
CO
^
CD
in
3-
ro
(XI
—
O
o
cc
LJ
Q.
M
O
^-
o
1—
cc
tncocncocD3-r-ino
ooooooooo
coro — cooo — coco
' r-in
orzcco-iino"
oi(XiiO3-ro3-ru i —
r cocorururu r~
0 — 0 0 — — — O
— co to ru ru ro r- i ru
r~ — in CD 01 01 01 —
0 — — OOOO —
* * * *
i o r> OD o ro in i r-
o co r- ru ru 3- 3-
— 0 0 — — — O
*
3- CD o to co 01 i i r~
r^ in — r^ 01 01 co
o — — o o o o
* * *
3- ru cn 3- 3- ro i i in
cn — Co Ol ru ru tO
0 — 0 0 — — O
coinrotororu i oin
oroocnoio coco
— — — oo— oo
* * * * *
to — r~(xiinco — o~
01 — o — ooruroco
^. .—.
* * •
co — • r^ 01 3- 01 oo — ^ ro
o — ooooo — —
* *
3- i in ru i t i o cn
0 — 0
*
t*) OD 10 if) OD ro to ^" j~
(Dcnror-r-coa> — CD
o — — oooo — o
* *
ru — ocoooor^CT)
— __oo — oo
* *
r-j-njoooo — CDOO
^_ _
*— * *-^ ^ *~~*
eotot~cn3-in i ruo
coroocnoo coo
0 — — 0 — — 0 —
* * * *
r~oo— i main —
cn tvi o o orocoai
o ~ — — — — oo
* * * *
cn in cn — i cocnairu
oiruoo o — in CD
o — — — — — oo
* * *
oODocooinrur^ —
cnrurocooo — cotD
0 — — 0 — — — 00
* *
^ ^ o_ m t— t~~ i~ >~
1 1 I 1 O O O <
r r o o o in in
GD CD 03 OD *~ *""
DD3DOOOII
NNMISIOOO<<
OOOO_I_I_IOO
ac.o-CLOococoocj
ozrzcccj-iincD —
o in
o ru
o ru
o ru
— 0
o cn
o ru
— o
o ro
o ro
— o
o cn
0 (XI
— o
0 —
o ru
— 0
o r^
0 —
— o
o cn
o —
o m
o 3-
— o
0 3-
— O
o ro
o 3-
— o
O CO
0 —
— o
0 to
0 (XI
*— '
o —
o (XI
— 0
o —
o ru
— 0
0 —
O (XI
— o
o ru
o (XI
MEAN (ALL SAMPLERS) 1
STANDARD DEVIATION 0
Lu
U-
O
CC
o
UJ
in
a:
0
cc
u
LJ
0_
CC
o
!r
LJ
h-
Q_
UJ
cc
LJ
REGRESSION VS Z INI
CO (XI CO
co CD in
Ul tjl UJ
000
— 3" Ol
— CD in
o ro o
ID — CO
— ro ru
3- oo r~
o o o
o o o
cn
CD ro cn
1_| Ul
Ol o —
0 (XI —
r- i~ r-
cn oo CD
3- 0 —
3- O CD
CO O ID
en r~ 3-
ru ru —
< r Q.
i i i
C/1 U. U.
O 1- 1-
ill
M N N
000
a. a. Q-
LJ LJ LJ
o z: z
— ru in
co in r-
CJJ CJ1 Ol
0 O 0
00 3- —
in cn o
cn r- ru
0 CO 3-
ru ru ru
in o co
o — o
000
01 3- 3-
— cn co
*— ' *— '
CO Ol 3-
3" CO 3"
(XI Ol 3-
r~ CD r~
3- — ru
r~ 3- ru
3- o in
in i— i—
I o O
cn o o
o — . —
z: a a
m
D O O
N O O
Q _J _l
< -I _l
a. co oo
LJ _) _l
or u _i
ro o co
OD
000
— — cn
in CD in
in ru in
in r- ru
ru ru to
r^ CD 01
o o o
o o o
3- o CD
ro 3- 01
cn 10 ru
3- 01 cn
ru ru ro
CO CO CO
ru o co
in r- cn
o o ro
03- —
t- >•
O <
0 O D
o in in
_i _J
Oil
o < <
_l C_> O
_i a o
CO O O
_J 3 3
cn o —
-138-
-------
p-
p-
co
CO
t
1C
1
ERIODS i
CL
cc"
UJ
£
CO
3
O
in
r
or
§
z
"
cc
CO
8
ct
-J
0
>-
§
t/1
§
in
cc
-
' u. u. in ° ° ° <
oi-i-o — — — z u.
z z o o o in in
>->->->- _j _i
CO 03 00 CO — —
NNP^tslOOO<<
OOOQ_J_J_JOO
CLCLQ-O-CDCOCOOO
UJUJUJUJ_I_I_I33
cz:zrro_ii/io~
PO
ru
ID
m ID
to nj
ru p-
(D —
in PO
10 ru
1C CO
in —
03 0
in ru
PR
10 CO
ID —
co co
P- O
ID — '
P- CO
in —
in 10
ru PO
CO 3"
10 ru
CD 0
lo ru
PO t
in ru
in —
p- ru
ru 3-
f ru
O
o —
t— LJ
I 0
3 Q
•z, <
< l-
ui in
N4
in in
in ru
in ru
10 ru
in ru
0 —
ID ru
CO J-
p- 3-
PO ru
ID ru
— ru
ru —
CO O
r"
in -
ru PO
co ru
in ru
o o
p- ru
in in
j- fy
o ru
co ru
— ru
3- (XJ
0
h-
—
— Ul
- o
• o
T. CC
O Q
Z <
< 1-
UJ CO
X
>
o
0
o
ID
in
3-
ru
-
o
—
co
CO
p-
1C
in
3"
PO
ru
„
o
o
cc
Ul
Q.
txl
O
h-
O
a.
P-O3-PO — oopoporu
— ruru — — — in 3-
in3-p- — 3-rucor^co
ooooooooo
oooru — ocoioru —
cnicp-03POPO3-pocn
oooo — — — oo
ouzccu-jino —
OOOO — — — o
(C I p- 1C o CO PO I —
oo co co ru ru PO ID
O o o — — — O
i p- r- o in ID in i co
ID 10 P- ru PO 3- oo
o o o — — — o
— in co r~ ru — i i in
co P- P- co j- in P-
OOOO — — 0
• *
— in cn o cn in i i o
CO 10 CO ID PO 3- ru
oooo — — —
j cn PO — • — 3- j co ru
PO 3- co in in PO ru
o o o — — o —
•
i icruporu3-r~oio
p-cor-rupnrorup-
*
— ru ru 3- ID P- P- i —
o 3- 3- r- o — ru cn
— o o o — — — —
• *
i i ru — i i i IP-
cn cn —
0 O —
cn to ic oo *~ PO m PO oo
OOOO — — — OO
* *
I P03-in3-— IDP-Cn
cop-p-rupopo — p-
ooo — — — oo
COP- coruru3-iom
OO O ~~ •— — ~ O O
co in ID ic co in i i co
CO CO P* P- PO 3- P-
oooo—— o
* *
i 3- o — i cn — ID 10
p- CD co po r^ PO 3-
000 — — 0 0
3-COP-p- 1 IDOinoO
cop-ooiD ruru — PO
oooo — — o —
*
PO p- o ru i co ru i o
ru in T m in oo m
— o o o — — o
1 1 1 1 O O O <
i/iii_U-inoooO3
Ql-l-O — • £u_
r r o o o in m
>->>->• _i _i
CO 03 CO 03 — ~
3333OOOIX
Nrsjp\irNJOOO<<
QOQO_I_I_ICJO
<<<<_!_! _JQQ
0.0.0.0.00000300
UIUIUJUJ_J_J_|33
Q£ZCCO_llfiO —
0 ID
o PO
— 0
o ru
0 3-
— 0
0 p-
o ru
— o
o in
O PO
— 0
0 PO
o pn
— o
o in
o PO
— o
o —
o in
— 0
O CO
o ru
o cn
o 3-
— o
o in
. ~~.
— o
O PO
0 PO
— 0
O CO
o ru
— o
0 3-
O PO
— CD
o o
0 PO
— 0
0 ID
.
— o
0 00
o ru
— 0
0 ID
o in
— o
•z.
o
t—
3 —
t- Ul
3 O
So
1- CE
3 O
— Z
Z <
< t-
ui in
u. co
U. 3"
o
O 0
g
o
• 3"
> in
win
?p-
1/1
r
cc
CC [^
o in
cc ~
UJ O
LJ nj
0 P
o ~
in o
SID
in
cc ru
u cn
t- PO
Q. 0
U (O
cc cn
UJ
Y-
—
<
1
01
V
in <
3
Z N
0 0
in
in <
UJ CL
or LJ
o
Ul
CC Q
p- co ru
ru p- ru
CO CO CO
O 0 0
ru o 03
o PO co
3- co in
03 CO ID
ru o o
o ru co
o o o
PO r- 3-
p- cn —
CO p- p-
o o o
ru in ru
§0 P-
O PO
(D 03 in
co co co
co ru UD
co co in
o PO in
— i
r Q. in
i i i
u, u. in
I- I- Q
>- >- V
< < <
m CQ m
hsl Psl fSj
O O O
< < <
O_ CL 0,
Ul UJ Ul
n z a
(XI (O
ID P-
tj) ol
0 O
in (O
cn 3-
co in
CO (D
in o
— 03
— o
0 0
ic co
p- 03
ru ru
03 PO
SPO
PO
p- in
PO O
3- —
PO co
ru in
i— t—
o o
o o
o o
88
CO 03
_l _J
0 _l
3- m
1C CD
CO ID
O 0 0
ru PO
in —
PO ru
cn o cn
iS PO
— ru
o o o
po ru
(£ ru
(XJ 03
— 0 O
O PO
in 3-
00 (D
p- o in
(D P-
p- P-
ru PO
— O PO
t- >-
O <
O O 3
— r u.
o in in
—i -j
i _j
Oil
o < <
_l 0 O
_l Q Q
CO O U
-J 3 D
in o —
-139-
-------
r-
cn
CO
""
E
ID
CC
cc
UJ
t-
o
m
5
-*
5
cc
z
I
*-
2
Hi
1
z
£
I
U
cc
UJ
z
l/l
>
-
in
3-
ro
*
-
-
CD
CO
in
3-
ro
ni
~
8
CC
l/l
-J
3
arzcru^oi-,,^
in — — — cn — ru ro r^ j- ro CD
— — — — —
SCO ru ro cn o r- i — in o co ro
co co co m ru in cn ro co j- j-
r^r-cor^cnoo to co o ru ru
i co r* o ru co J* i cn CD o co r-
ru r^ co CD ro ro cn cn CD o —
— — 01 j in in CD HJ ru j ro
ID — nj r- r- ro i icoj-ojcn
co o ro j o o ru j — ru o
CD o o co ru ru cDoromin
— — ID o Oi cn i i ruiDooro
in (xt in o o 01 r- — cn ru in
Sj-rucor^m i roorooruo
ru cn j- cn 01 ru — ro j- 3- cn
coScneooooPPcoScncn
ipiDj-ococn — roiDOjomo
ryr-r-j-— rufoiDr-m — j-o
roruruojfOfororurunjromin
i i
i icor-i i i r- j- ID o ro ru
fO— ru — ro ro ro j-
— eniDiDinr-— iDcoooruo
r-eoeotococncninior-cnoo
iDorucncooin — j-moj-
tno — comcoroocnj-oino
cn— ocnnjrufoocncnru — ro
ro o cn — ro r- i iDooococo
in ru — ro r- r- ID r- ID r- o cn
cn GO CD — ru ID CD r- o
rur-ruruitncnmocnoi i
inmcomiro— ifoiooi i
J- in r*- j- cn ro in co o
CD cn CD r- o— r-ioo
I I I I O O O < U. O
o »- t- o — — — nt-u,— ui in
>>->->- _J _l _l ZZ
UJLJUJUJ-J_l_lI)DDUJU.U.
CO
ID
ID
2
$ -
— in
in co
cn —
ID J-
ru ru
S|
—
ru in
ID r
ro ro
a-
ID ro
ru cn
ro
in —
2 "
— j-
in in
ss
ro CO
ro ro
— —
o cn
r~ o
~
ru T
ru r-
cn —
as
CO —
CO CO
cn in
CO —
-|
UJ <
|2,
~ Z
Z <
uj in
cn 2
CTl ID
ru r-
m —
cn CD
—
co ro
3- ru
8i -
ru 3-
3- ro
5 ro
CD r-
CD -
r- cn
ID (XI
ift r-
— —
ss
~
J- 01
cn —
™ S
cn nj
ID 10
— in
- z
*: o
CL t-
H CC
e'l
z <
LJ l/l
r
LT!
>
-
in
j-
ro
XJ
~
-
cn
CD
r-
in
j-
ro
ru
"
o
o
or
ISl
O
O
^T
oooooooooooru —
ooooooooooooo
cocncnco — — — corjocn — — ru
OOOO — — — OOO — — —
OrZCCU-JLOOI — Q_->^
CD 01 0! CO — — — • COO1OOO
OOOO — — — O O — — —
SSSSSS^ ' PSiss
OOOO — — — O O — — —
i CD ru r- to — o i oruooro
CD cn r^ — ru ru r- o o — o
• •
• • *
r- CD CM r* o o r^cn — roro
OOOO 00
*
o co — to ro ->->-> -J_1_J ZZ
<<.<.<. _j_i_jcnoo
ODOP03CQ ™ — "- LU LO ui
DDD3OOOIIIOQQ
QOQQ.J-J-JOUOCCII
UJUJUJU'_I_J_)DDZ)UJU.U.
o -~
— o
o —
— o
o cn
o —
— 0
o r-
o —
• •
o ru
— o
o ro
o —
O t\j
o ru
o — .
— 0
0 CD
o ru
— 0
O CO
— 0
o CO
o ~
— 0
o —
o ru
o —
• •
— 0
O IO
— o
O CD
— 0
o —
o ru
— o
o CD
o —
— o
-1
or ^-
UJ <
ll
- Z
z <
< 1-
LJ in
r
ucncnmcncncncncncncnmcocn
o ....
OOOOOOOOOOOOOO
a:
O
(J
• fOicn^-rucnor-cnr-coj-roin
>r^cocDo — j-cn — rurooo —
ujoj- — — ruj-foj-iDinj-ruru
-j-inoomrucDCDru — rucn —
tnj-cniDinroinincnp*ocoiDLn
or
gcnmruroru(Dr-cnroco — ruiD
j-oomj-roj-incoiDcor-inr'o
CCOOOOOOOOOOO — —
cc - • • •
LJcniooroj-corucorooicoCDnj
o_ — — j-u)oincD — mmiDinm
ocoooicoru — — cor^cnocn —
1/1O — OO-- — — OOO — O —
ccr-cncninruiDJiDj-r-iDcnru
or^rorur-coinioj-j-ruoin —
trinruiDcnjrioooinrucnj-cn
Q; . . - ....
ujofur-r--LOooiocnr-(D.3-m
mcninj-roinmoiocnr-r-in
I— — — incoLOcncnco — j-ro — —
CLinoiruoocnrocoiotDiDiocD
LJ — ror~ino — ronjODroinotD
o •
oroo — ruromruoitor-j-innj
LJ 10 .r — ro cn — i — in in ro — m
t-i iii i i ru
z
<2IQ-Lni-K-t->- _l fXI J-
I I I I O O O < U. O • •
SLtU,LnOOUD_ir3>
*- >- a — — — Et-u. — oic/i
r ZOQQLO»-1/1I_I_)
Ol<<<< _J-J_IU1OO
ODD^OOOillQQQ
ZMMrsjrsiOOO<<
-------
iPJl-jOLPfOj-J-OlOltO
r>
01
CO
1
(O
1
" '
1
a.
cc
u
CO
CO
U1
kNOGRAt-
z
-
a.
^
*?
u_
o
Q
f-
cc
a.
o
cc
LJ
z
S
0
CO
in
3-
ro
ru
-
o
cn
03
r-
co
in
-
ro
ru
—
RESULTS PERIOD
o x:
ru cn
nj in
r- ro
ro r-
i cn
in cn
in to
cn cn
cn 10
in r~
r- 10
ro cn
cn cn
3- ru
in i
to IP
ro oo
1 8
— CD
o cn
IP O)
IP O)
fO CO
r- to
r- nj
r- —
(0 GO
< r
i i
z:
00 03
Nl N
Q Q
0. 0.
LJ LJ
O £
Z CC
CO 0
ru to
cn ID
to to
Ol CO
ro r-
r- nj
00 CD
0 3-
in —
— to
3- in
3- cn
ru cn
ru ru
ID 3-
10 in
O 0
— r-
3- cn
CO 10
ro r-
— GO
CO 0
o r-
ru in
cn —
r- ID
0- (ft
i i
b. LO
"8
V V
CO CD
NJ N)
Q a
0. CL
LJ LJ
Z CC
0 _l
— in
m m
— in
ru —
r- r>
r- cn
o co
cn r-
ru in
ru cn
j- t>
ro nj
. i •
ru ru
CD ID
in 3-
3 3-
1 I
ro in
r- -
r- ro
cn 3
en CD
WE
i cn
1 — •
O
r- ru
3 S
r- r~
h- h-
O O
u u
o o
88
CO CD
r- o
ru o
oo ro
in r-
h—
LL. —
LO Q
-I CO
< O
cj cc
Q <
U 0.
O LJ
— O
cn
CO
ru
in in
ID r-
ru r-
00 3
ID ro
3- ro
o —
cn r-
ro ru
r- —
ID —
in o
ro —
o cn
ru in
— 01
ro —
CD CO
in 3
— to
ID ro
ru —
co r-
ro r~
ro cn
— r-
ro o
cn cn
ru ru
o co
r-
2 ME AN (ALL SAMPLERS)
STANDARD DEVIATION
ro
r
CO
r-
3-
O
3
ID
u)
to
^
-
in
ro
CO
J-
o
ID
f-
ro
cc
0
3"
ru
CO
cn
CO
CD
o
o
r
a
z
LJ
3:
X
CD
3-
3"
ro
o
3-
fi
CD
"~
CD
CD
o
ru
ro
CD
LO
0
CO
CD
in
LO
in
CD
cn
ai
STANDARD DEVIATION
O
C/l
O
CD
in
3-
ro
ru
-
0
01
CO
r^
CO
LO
3-
ro
ru
—
o
o
cc
LJ
CL
Nl
0
o
cc
incnioin3-in3-cocoLO
ooooooooo —
oooooooooo
r-inrucncococo — roo
cnoocoooocncno
o — — o — — — oo —
Q3IZCCU_IC/1O— O
3-oiLOiDr-cor- i coin
cncncnooooo co —
o o o o — — — o —
ro o) — cn CD in co i ror-
cncnocoooo cno
o o — o — — • — o —
i — in o ro 3- 3- i — ro
cn cn co — — — ocn
ooo — — — — o
* *
orucnoioini ico —
oioicncooo cnro
oooo — — o —
* * * *
ro o — ID co ro 3- i inro
cncncncoooo onj
* » • *
3-ro — coocnrur-oco
cn — oco — o — cooco
* * * *
ru3-ro — cntoiO3-r-o
cn — ocnooocncncn
cocnr-coiDcoru3-oro
ocncncoooocncn —
— ooo — — — oo —
o i r- r~ i i i co co in
— — o — — o
* * * *
ro — ruco3-LQOD — oco
o — — coooococncn
— — — o — — — ooo
i ruru — r-3-j- — — o
— ocnooocnco —
* * * * * *
ior>inrur-inr>cnrocn
cnoocnooocnor>
i_i ^j u —
cnj-inrucoto i oocno
o — — o — — ooo
* » » »
ruooj- i m3-coinfo
— — — o — — ooo
* * *
r>oincn i (DinocDru
01 — OCO — — 03 CD o
o — — o — — o o —
» * *
C~COrOCO(OL03-3-rO3-
cn — — ODOOOCOCDO
• * * * +
I 1 I 1 O O O < O
cnu-u-LnooooDo
z: roBocnoiQ
QD CD QQ QD -— "— LJ
DDD30OOIIO
MNJMNOOO<DLJ
O —
O —
— 0
0 0
o —
— o
o r~
0 0
— 0
o ro
0 —
— 0
o in
o —
— o
o ru
o —
o —
o —
O CO
0 O
0 CO
o o
— o
o co
— o
0 0
— 0
0 O
o —
o cn
O 0
*— '
o in
o —
— o
o —
— o
o ro
0 —
— o
o ru
o —
ME AN (ALL SAMPLERS)
STANDARD DEVIATION t
u_oi — oruj-corurooDcn
u.coincococncocnr-3-ro
ooooooooooo
cc
O
Ljnjtpnjfijioioj-ioj-o
t/iroooiPJ-roj-roiocOiP
cc
o?cntp!.rro..r.±t^coio
ccooooooooo —
LJOOOOOOOOOO
LJ to r~ j~ CD r*- — • CD r"" o 01
o_njj-cocn — r-r-cor~ro
tJ U) Cj U) (JJ i*J i—> \J) UJ UJ
ccororofor>coipnjroco
ororor-iPCO — iP^roj-
Q_ — ro to •— • to ro — in ID ro
LdQicnrO'-'fxjcoooipru
o^r^njr^ oir*-oorot.oj'^-
LJ nj ru i ru IP nj j-
h- 111
<.ZQ_aii— i— i->- t-
t l l l O O O < O
t/)U_U_l/>OOCJO.D(_)
Ot— t—D — — — HU_ —
r z:aaai/iino
M
>• >• V >- _J _J
Ul<<<< _l_Jl/l
>CDCQCDCQ ™ — . LJ
.D-DD-DOOOIXO
ZNINMrMOOO<
-------
p-
OD
Z
ID
CC
or"
5
O
3
o
kNOGRAMS
z
••
CD
0.
LJ
L.
O
1
(—
5MPARISON J
1 INTERCC
o
o
ID
in
j-
ro
ru
o
01
CD
p-
10
in
3-
ro
ru
8
oc
LJ
0.
in
h-
in
LJ
a:
arzcEo_iino —
j- o to o ro ru 3- i i o
in in CD P- CD o o
01 p- ru j ro ru i inj*
— in j- ro ID r- ruoi
ru — — ru j- j- — ru
airuoiruoDiominin
j-cDcnmp-— OP-ID
ID P* o — roininmin
i i LO in i i i ru j-
ru — — nj
j-iOLOininoocviffiru
ru
i rooroj-ocoocn
rooiin — iniooio
j-rup-oiiDj-o — p-
S — ID r — ru i o r-
ru — f f LO ru o
— ru ru — ~o ro — ru
i o o in i 03 in o ID
— 10 ru in i 3- ai ro ro
rurur-o top- — in
— ru — — ru ru — —
5ro in o i — CD i t
y- m ro mm o
• - ro M -
1 1 1 1 O O O <
Ot-t-O — — — Z LL.
r z; o o o in in
CD 03 CD CD — — •
OOOO_I_I_IOO
0.0.0.0.00000000
azzortj-jtno —
ID
r-
01 in
0 (D
ru
ao in
— p-
o —
ro p-
p* ro
O ID
ru —
(U —
0 00
ru
So
(U
r p-
— CD
0 P-
ro p-
01 —
ru -
10 in
ru oo
ru
ID y
ru —
en 01
P- ID
Z ME AN (ALL SAMPLERS) 197
STANDARD DEVIATION 125
o o
Ol ID
01 CD
LO 00
00 Ol
ru p-
r ru
ro co
PS
o cn
10 —
CO ID
ru p-
CD P-
ru "
10 00
ru 01
ru co
ru CD
ru
ID 3-
in p-
P- LO
X MEAN(D,M,L,G, I ) 192
STANDARD DEVIATION 110
0
o
0
LO
in
£•
ro
ru
o
Ol
CD
p-
10
in
ro
ru
o
o
C£
LJ
a.
rsi
O
o
t—
O
ozzox_)_iino —
o o
O f
— o
o —
o ro
— 0
O LO
o ro
— o
o ro
o ro
— o
o —
o ro
— o
o —
o ro
— o
o ru
o in
— 0
0 0
o j-
— o
o ro
o ro
— o
o on
— o
0 Ol
o ro
— o
0 ID
o j-
— o
o ro
0 J-
— 0
o ao
o ro
— o
O (D
— 0
o 01
o to
— 0
MEAN (ALL SAMPLERS) 1.00
STANDARD DEVIATION 0.63
u.
u.
6
o
0
i
o
LJ
%
ff!
or
LJ
t—
a.
LJ
REGRESSION VS Z INTERC
— ru r-
CD 00 P-
O 0 0
10 in 01
— — p.
ru T ro
ru oo LO
ooo
o — 01
ro p- OD
ooo
o CD cn
cn 01 in
OD ID P-
to o LO
ru j- ro
03 Ol in
SOD 0
i ro
< z a.
i i i
in u. u.
O 1- t-
z
>->•>•
M M M
OOO
a. a. o.
LJ LJ LJ
O Z Z
01 in io
ooo
ru ro ID
OD ru j-
fuioS!
oi in j-
o o o
3- r ru
ru 3- ID
o — —
to P- to
— ru (u
— in oo
ru ro ro
in ru p-
ru ru ru
ro —
i i
to i- i-
1 O O
tn o o
zoo
D O O
MOO
Q _l _l
•< _J _l
0. CD CD
LJ _l _l
or o _i
— ru ID
ro in J-
cn 3- p-
ooo
ro p- in
03 r~ 01
p- cn 01
j- in to
j- ro j-
01 m o
ooo
oo ID cn
ru (u t
— o o
o ro to
P- 00 ID
in CD t
3- in in
j- ro j-
ro r- 01
— ro 01
P- Ol -
oo ro 01
— in
i- >-
o <
003
— Z li.
o in in
855
_! O O
_l O O
CD O O
in o —
-142-
-------
TECHNICAL REPORT DATA
/rYc'JSi read />;j(7.-c nous en r/ic ii'nrsc be ton rt/";;
1 REPORT NO J2
EPA-600/7-78-118 [
4 TITLE AND SUBTITLE
INTERCOMPARISON OF SAMPLERS USED IN THE DETERMINATION OF
AEROSOL COMPOSITION
7 AUTHOR(S)
D.C. Camp, A.L. Van Lehn, and B.L. Loo
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Lawrence Livermore Laboratory
Livermore, CA 94550
Lawrence Berkeley Laboratory
Berkeley, CA 94720
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory-RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
.Vfm.t"
.3 RECIPIENT'S ACCESSION NO.
5 REPORT DATE
July 1978
6 PERFORMING ORGANIZATION CODE
8 PERFORMING ORGANIZATION REPORT N
10 PROGRAM ELEMENT NO
1NE625D EB-11 (FY-78)
11 CONTRACT GRANT NO.
IAG-D6-0800
IAG-D7-F1108
13. TYPE OF REPORT AND PERIOD COVERE
Final
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16 ABSTRACT
An intercomparison study was carried out to evaluate the performance of 11
different designs of aerosol samplers. The samplers were operated by participating
scientific groups having recognized expertise in sampler development and operation.
The devices tested include hi-vol, TWO MASS, cyclone, CHAMP, streaker, stacked filter
and manual and automated dichotomous samplers. The samplers were operated in Charlesti
WV for eight consecutive days duirng May of 1977. The collection surfaces of each
sampler were changed at least every 12 hours which enabled the intercomparison to be
made for 16 sampling periods. The collected samples were returned to the laboratory
of each participant and analyzed for mass, nitrate, sulfur or sulfate, lead, and 9
other elements. Most of the samplers separated the aerosol into two fractions with
50% separation diameters ranging from 2.4 urn to 4.3 urn. The upper 50% cutoff diamete
for the various samplers ranged from 14 um to about 30 urn. Best agreement among
samplers was found for elmeents such as sulfur and lead that occurred primarily in
the fine fraction. The amount of total mass collected was strongly influenced by the
upper 50% cutoff diameter of each sampler. For stacked filter samplers and the tandem
filter samplers, the fine fraction appeared to be enriched with crustal elements such
as Si, Ca, and Fe, which suggests that there are particle bounce errors. Of all the
samplers tested, the automatic dichotomous sampler showed the greatest precision.
17. KEY WORDS AND DOCUMENT ANALYSIS
a DESCRIPTORS
*Air pollution
*Aerosols
^Samplers
*Comparison
*Mass
*Sulfur
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b IDENTIFIERS/OPEN ENDED TERMS
19 SECURITY CLASS (This Report)
IINfl ASSTF.TFD
20 SECURITY CLASS (Tins page)
UNCLASSIFIED
C. COSATI
Held/Group
13B
07D
14B
07B
21. NO. OF
151
PAGES
22. PRICE
EPA Forrr 2220-1 (Rev, 4-77) PMFVIOUS EL i T'C-J > s -, BSC LE-- TE
143
-------
W
O
O
OB
I
03
8 "S ^ 8? x
•• f> c; £•
"•*- -*. "t VJ
*+* 5 s -
>, CD ^>
•2 •-* j
l>l 2* Q rj
Do o
o 3^
Oj
S. QJ -4
=;• o- s
S ° -
Q <* (t) Ct
^5 ^s;
li.ln
J» 5
a ~»
Q. »
•-.
* S
«^,. Cft
I?
o o
=.. si
Q
a
Q>
O-
o
Q)
O-
m ™
c>
ti*
O -n -n
•D 0 o
-V X z-
R
PR
H * 00
C < C
Z > W
~^ m ^
< ._ m
™ m V
% »
2
T)
r
O co
< S
m °
c
CO
51 O rn
-• - <
O 5:
(0 J3
o O
£.
O 1
5' S
O 03.
1 »
5 ro
to (/)
±- CD
QJ
i*
O 3^
*• ?
01 1
ro
a
00 o
3
o
CD
r-r
CD
«t> m
w 2
•1
o r-
to x
ag
O m
CD o
• S
o O
TD Z
1 >
^ O
5- m
z
o
m
------- |