vvEPA
United States
Environmental Protection
Agency
Environmental Sciences Research EPA-600-4-78-037
Laboratory June 1978
Research Triangle Park NC 2771 1
Research and Development
Dispersion of
Pollutants Near
Highways
Experimental Design
and Data
Acquisition Procedures
-------
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 ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors,
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/4-78-037
June 1978
DISPERSION OF POLLUTANTS NEAR HIGHWAYS
Experimental Design and Data Acquisition Procedures
by
S. Trivikraraa Rao, Marsden Chen, Michael Keenan, Gopal Sistla,
Ramam Peddada, Gregory Wotzak, and Nicholas Kolak
Division of Air Resources
New York State Department of Environmental Conservation
Albany, N.Y. 12233
R-803881-01 and R-804S79-01
Project Officer
William B. Petersen
Meteorology $ Assessment Division
Environmental Sciences Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, N.C. 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U. S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
The major thrust of this investigation was the collection of data on
particulate and gaseous pollutant concentrations, and detailed micrometeoro-
logical data in the vicinity of a major highway in a non-urban setting. A
site on a relatively undeveloped section of the heavily travelled Long Island
Expressway was selected for the collection of this data for the purpose of:
(i) documentation of the distribution of sulfate, lead, total particulates
and carbon monoxide at an array of sampling points adjacent to the highway;
(ii) studying the micrometeorology associated with the highway, with special
attention to those parameters important in the determination of atmospheric
dispersion; (iii) reevaluating highway air pollutant emission factors from
data gathered in tracer gas experiments; and (iv) examining the applicability
of existing highway air pollutant dispersion models.
The location of the site and the experimental setup for collection of
pollutant data are described. Details of the data acquisition procedures
also are presented in this report.
111
-------
CONTENTS
Abstract iii
Figures vii
Tables viii
Acknowledgments 1x
1. Introduction 1
2. Site Description and Initial Preparation 4
3. Carbon Monoxide Data Collection 10
General Set-Up 10
NDIR Description 11
NDIR Modifications 14
Averaging Chambers 15
4. Particulate Data Collection 18
Description of the Sampler 18
Sampling Procedure 24
5. Traffic Data Collection 27
6. Meteorological Measurements 29
Climet Instruments 29
Gill UVW Instruments 30
Other Instrumentation 31
-------
7. Tracer Gas Experiments 33
Field Set-Up 33
Sulfur Hexafluoride Analysis .36
8. Special Micrometeorological Experiments 41
Instrumentation 41
Data Processing 43
9. Computer Data Acquisition and Reduction Techniques 44
Computer System 44
Computer Programs 46
10. Laboratory Analysis 51
Particulate Weights 51
X-ray Fluorescence Spectrometry .51
Anion Analysis by Ion Chromatography 52
References 55
-------
FIGURES
Number
1. ROAD project site locale 5
2. Initial sampling set-up 8
3. Modified sampling set-up 9
4. Modified CO sampling set-up 12
5. Functional diagram of a NDIR 13
6. CO sampling system with the vertical profile system 16
7. The flow diagram for the vertical profile system 17
8. Dichotomous sampler location in the sampling plane 19
9. Dichotomous sampler flow pattern 20
10. % of aersol concentration and overall losses 22
11. % of aerosol concentration for different flow rates 23
12. Sampling locations for SF& 34
13. SF6 sample collection 35
14. Schematic of SFg gas release set-up 37
15. Route for the tracer gas release 38
16. Location of the sonic head 42
17. Configuration of the computer and its accessories for
data collection 45
18. Typical computer output 49
19. Configuration of the computer and its accessories for
data analysis 50
20. Ion chromatograph flow system 53
vii
-------
TABLES
Number
1 Duration of Data Collection 3
2 Instruments Used in the Project ..... ... 6
3 Log Sheet Used for the Dichotomous Sampler 25
4 Length and Speed Categories 28
5 Log Sheet for SFg Tracer Study 39
6 Arrangement of Data Within a Given Block 48
vm
-------
ACKNOWLEDGEMENTS
The authors are grateful to many staff members of the Division of Air
Resources for taking part in the tracer gas release experiments. Special
thanks are extended to Messrs. John Hawley, David Romano, Gerard E. Blanchard,
William Frenz, John Hawkins, Gary McPherson, and Stanley House for their help
in conducting this study.
The cooperation of Mr. Ronald Piracci of the New York State Department
of Transportation and Dr. Ulrich Czapski of the State University of New York
at Albany is gratefully acknowledged.
IX
-------
SECTION 1
INTRODUCTION
The Environmental Protection Agency has developed national ambient air
quality standards and oversees State enforcement of these standards. Before
a permit is issued to construct a new highway or make major modifications to
an existing highway, an air quality assessment must be made to ensure that
the resulting vehicle emissions will not cause a violation of the national
ambient air quality standards. In order to predict the air quality impact of
motor vehicle traffic, mathematical modeling techniques are being employed in
various forms to estimate ambient concentrations resulting from these vehic-
ular emissions. A new national concern has lately arisen adding sulfate
particulates from catalytic converter emission controls to the list of
automotive pollutants.
In order to test the ability of any mathematical model to simulate the
advective-diffusive transport of gaseous or particulate matter in the atmos-
phere, an extensive data base of meteorological conditions and highway-related
pollutant concentrations must be collected. While there have been a number
of mathematical models developed for the prediction of pollutant levels, only
a few experimental programs have been undertaken for establishing a suffi-
ciently detailed data base to be used for model verification. The objective
of this investigation was the collection of such a data base at a highway
site on the Long Island Expressway. This site has an average annual daily
traffic count of 100,000 vehicles/day, and a relatively low background con-
centration of pollutants from sources other than the highway in question.
Data on carbon monoxide, sulfate, lead and total particulates, traffic counts
and meteorological variables were collected during the period October 1976
through May 1977.
This research was intended to cover all aspects of pollutant dispersion
in the vicinity of a major highway. This report describes the experimental
setup and design and the data acquisition and reduction techniques. The
location of the site and preparations for the data collection are presented
in Section 2. The carbon monoxide sampling procedures are described in
Section 3. The ambient aerosol data collection is presented in Section 4.
Techniques of measuring the traffic according to vehicle length and speed are
described in Section 5. The various meteorological instrumentation used and
the data collection procedures are presented in Section 6. In order to test
the estimated pollutant emission rates, tracer gas release experiments were
conducted and the details of these experiments are given in Section 7.
Special micrometeorological measurements to study the turbulence character-
istics in the vicinity of the highway are presented in Section 8. The
-------
logistics of the computer are presented in Section 9. The procedures of the
laboratory analysis of the particulate filters are described in Section 10.
Table 1 lists the various parameters measured at the site. All the data are
being analyzed to examine the validity of various assumptions underlying
mathematical modeling of highway dispersion and to determine the simulation
capability of various mathematical dispersion models. The results of these
studies will be presented in a subsequent report.
-------
TABLE 1. DURATION OF DATA COLLECTION
Measurement
Sulfur Hexafluoride
Particulates
Carbon Monoxide
Carbon Monoxide
Traffic Counts
Traffic Films
Gill UVW
Climets
Relative Humidity
Precipitation
Solar Radiation
Temperature
Temperature Gradient
Sonic Anemometer
Temperature Fluctuation
Wind Profile A
Wind Profile B
Sampling
Length
1 hour
2 hour
10 min.
10 min.
10 min.
4 hour
10 min.
10 min.
10 min.
10 min.
10 min.
10 min.
10 min.
1 hour
1 hour
15 min.
15 min.
Duration
23 runs
236 runs
3 months
2 months
8 months
32 films
8 months
8 months
8 months
8 months
8 months
8 months
8 months
36 runs
15 runs
500 runs
250 runs
Number of
Sample Location
16
8
8
11
6 lanes
1 direction
4
6
1
1
1
1
1
1
1
6
4
-------
SECTION 2
SITE DESCRIPTION AND INITIAL PREPARATION
The sampling plane (Fig. 1) was located several hundred meters east of
Pinelawn Road near the Exit 49 interchange of the Long Island Expressway
(Interstate Route 495). The road bed, oriented approximately east-west, is
level for about one kilometer on either side of the sampling plane. The high-
way has 3 lanes in each direction with a 20m wide grass median. On the
northern side, the shoulder bed rises gradually to a height of about 2m above
grade at a distance of 21m from the road edge, beyond is a sod farm having a
flat unobstructed fetch. Similar conditions are found on the southern side,
except for a small grove of trees located directly 200m south and a few houses
500m southeast of the sampling plane. There are some commercial establish-
ments to the west of the site about one kilometer away.
Since the site was within the Expressway's right-of-way, permission was
given by the New York State Department of Transportation (NYS DOT) to clear
the immediate environs of shrubs and small trees. This was done for 15m on
either side of the sampling plane. From Pinelawn Road, NYS DOT constructed
an access road about 100m south of the highway. Clinker from boilers at a
nearby power station was used to provide an area for the trailer and a
parking lot. This surface helped to control vegetative growth in the summer
and provided traction during the winter months.
Coordinates of the area, as provided by the Federal Aeronautics Adminis-
tration, are 40°47'-15" N, 73°24!-30" W, and the elevation is 66m above the
mean sea level. No obstruction, lights or special painting were required
for the towers.
Most construction and erection of sampling equipment was done by the
Division of Air Resources personnel, A 70 ft x 40 ft x 10 ft high chain link
fence was constructed in November, 1975, to provide security for a refurbished
40 ft long tractor trailer modified to include electrical power, heat and air
conditioning, telephone service, work benches, instrument shelves and office
space. Prior to obtaining the trailer, free of charge, the budget allowed
for two aluminum sheds. The benefit of hindsight shows that such a course
would have created very difficult working conditions. Table 2 lists the
equipment used for the CO and meteorological data collection.
In January, 1976, the NYS DOT with its own rig and crew attempted to
drive guy-wire anchors at the project site. However, snow, rain and thaws at
different times thwarted these efforts. The final anchor was placed during
the first week of March, 1976.
-------
™* B Afi^HMt ttrJIn^V ''
t #*Dmvv oouww \
^5^°^ X I aonqoon Scte^
J; ROAD PROJECT SITE LOCALE
-------
TABLE 2. INSTRUMENTS USED IN THE PROJECT
Instrument
Beckman 865
Climet 011-1
Climet 012-1
Climet 012-10
Gill Anemometers
Spectrosun
Pyranometer
MRI 840-1,840-2
MRI 840-2
MRI 840-7
MRI 302
Measurement
CO
Wind Speed
Wind Direction
Wind Direction
U, V W Wind
Sun and Sky
Radiation
A T
T
Rel. Humidity
Precipitation
Principle
Detection
Range
NDIR 0-50 ppm
Anemometer Cup 0-20 m/s
Vane 0-540°
Vane 0-540°
Propeller 0-20 m/s
Thermopile 0-2 Langleys
Thermistor -3°C to +3°C
Therm./Resistor -30°C to +50°C
Crystals 0-100%
Tipping Bucket 0.01-1"
Output To
Computer
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
0-5 VDC
-------
The following is a list of material used in erection of the towers:
1. Trylon, Inc. (defunct) STT-650H-40, STT-650H-80 tower sections. Each
10 ft section consisted of 3 (3/8") rods tied by a lattice of 1/8" rods to
form an equilateral triangle cross section of 7 1/2" side. These are spot
welded with tabs at either end for bolting.
2. Three telescopic towers mounted on trailers could be cranked up to a
height of 65 feet. These towers were of sturdier design than those described
in 1. Manufacturer unknown.
3. A third unmounted telescopic tower (70 ft), which proved to be the
strongest, was used as the median tower. Manufacturer also unknown.
The three mounted towers were on loan from the U.S. Environmental Pro-
tection Agency (USEPA) and the others were from the State University of New
York at Albany (SUNYA).
4. A.B. Chance Co. PISA-5 8-inch helix anchors with 3/4" rod driven to
a depth of 5 to 10 ft, depending on number of wires attached.
5. 3/8" standard galvanized steel wire cable for guy wires, turn buckles
and clips. 3/16" wire was also used to support the median towers horizontally
to the shoulder ones.
6. 3' x 3' x 1/4" hot roll steel plates as bases for unmounted towers
and two lightning rods for the 25 meter towers.
7. Climbing rope, harnesses and boots.
Placing of the median tower required help from the Suffolk County Police
Department, Although the guy wires parallel to the longitudinal axis of the
highway were tied to anchors in the median strip, those in the transverse
plane posed a slight problem. The police were required to halt traffic in all
three lanes on one side of the highway, until 3/16" guy wires could be strung
from the already erect median tower and tied to the ones on the northern and
southern shoulders. Three sets of wires on either side were placed at heights
ranging from 25 ft to 45 feet. During this time, the median tower was held in
the vertical position by means of a crane from NYS DOT. Figure 2 shows the
initial setup at the site.
Tower 12 suffered structural damage when Hurricane Belle struck Long
Island in August, 1976. The upper 16m sections were later removed for safety
reasons. Therefore, the temperature difference measuring system was trans-
ferred to Tower 11. This new setup is shown in Fig. 3.
-------
00
E(25M)
D (16M)
C(8M)
B(4M)
A(2M) (
C
0 (
c
CI-3
f ^
) (
f [
p
)
)
f C
O CO VWT°
® CO a UVW A AT
*T° CLIMET D PARTICULATE
T SF, ^
CI-3
*?
c
CI-3
A
?
I C
O
'
CI-3
p ^
|Hl»Nw«y|
'
; j
CI-3
7
1
1-25
^
'
I C
0 (
E
p c
r <
CI-18
f 1
P
)
)
1 f
456
MMPLIN* POINTS SUBJECT TO CHAN8E
SCALE)
8 9 10 11 12 13
TRAILER
Fig. 2 Initial sampling Mt-up
-------
O co
T SFs
wv °T
A AT
07"° Climet Wind Speed 8 Direction
X Gill UVW Anemometer
Q Dichotomous Particulate Sampler
DC16M)
C(8M)
BHM)
A (2M)
CI-S
I6M-
2IM
456
Fig. 3 Modified sampling setup
13
Not to scale
-------
SECTION 3
CARBON MONOXIDE DATA COLLECTION
GENERAL SET-UP
Teflon tubes, 7/16" inner diameter x 1/2" outer diameter, were strung
from each assigned point in the array, shown in Figure 2, back to bulkheads in
the wall of the trailer. Each tube continued from the bulkheads and termi-
nated at one of three cylindrical collectors. All tubes between the bulkhead
and collector included a Brooks Instrument Model #1355-8900-21-C-1-B-AA
Sho-Rate "150" flowmeter, and a "T" section from which a sample could be
diverted to a CO analyzer. Diversion of the sample was accomplished by means
of electrically activated 2-way solenoid valves which were energized by
signals from a mini-computer.
Each collector consisted of a 9" diameter x 2' long enclosed steel
cylinder with 12 couplings on top at the periphery for the teflon tubes.
Another coupling was placed in the center from which a 1/2" copper tube led
to a high capacity pump. Internally, a 1" pipe went from the center coupling
to 1" from the bottom of the cylinder. When in operation, the collector was
exhausted by a high capacity pump via the copper tube and air from the
sampling points entered continuously via the teflon lines. The purpose of
the collectors was simply to maintain sample flows with only three pumps.
Each of the 32 teflon lines was fitted at its sampling point with a
polyethylene funnel and mesh for water and insect protection. Within the
trailer, the lines were divided into 8 groups of four. Sample flow to one
of eight Beckman Model 865 non-dispersive infrared (NDIR) analyzers from each
teflon tube was controlled by the normally closed solenoid valve referred to
above. The on-board mini-computer was programmed to energize 8 solenoids
(one from each group of 4) for 75 sec., allowing the sample from each of
these 8 lines to be analyzed by the corresponding NDIR. At the end of this
time the valves were deenergized and therefore returned to their normally
closed position, thereby inhibiting any flow from those sample lines to the
NDIR analyzers. Simultaneously, another 8 valves were energized by the
computer for 75 sec. and analysis of these samples occurred. This procedure
was repeated for the third and fourth valves in each group to complete
analysis of all 32 lines. Since polling was done on a continuing basis, the
computer then switched back to the first valve to repeat the cycle. For a
ten-minute period, the computer routine averaged the first and fifth reading,
second and sixth, etc. Thus, for each NDIR, the ten-minute printout had four
values.
10
-------
A change in the sampling set up was made in December, 1976 after tests
showed-that the 2-way solenoid valve system was inadequate. This was proven
by introducing a known concentration of CO gas at a sampling point on the
tower. Only 80% of that concentration was recorded using the 2-way valve
system, which was allowing a back-flow from the collector to dilute the
sample line flow. To overcome this problem, it was decided to convert from
the 32 point sampling array to an 8 point array, which included points at the
2m level, (3A through 10A inclusive, see Fig. 4) or 1 point per analyzer.
This was done with the NDIR analyzer pump sampling directly from the line.
It was determined that the basic, 32-point, system could have been
retained by changing from 2-way to 3-way solenoid valves. However, the
decision was made not to incorporate 3-way solenoid valves for the 32 sampling
points because a deadline of March 31, 1977 to end the project had been agreed
upon with the NYS DOT.
NDIR DESCRIPTION
The Beckman Model 865 non-dispersive infrared (NDIR) analyzer was used
for measuring the carbon monoxide (CO) concentrations. There are two sources
of infrared radiation as shown in Fig. 5. An optical chopper is placed
directly in the path of radiation and interrupts it at 10 HZ. When in
operation, a portion of the infrared energy is absorbed by carbon monoxide,
with the percentage of infrared radiation absorption being proportional to the
concentration of carbon monoxide. A flowing reference cell is incorporated
in the analyzer and is designed to minimize the effect of water vapor inter-
ference and to eliminate the need for bottled zero gas. A catalyst scrubber
is also placed upstream of the reference cell to remove carbon monoxide; the
sample cell receives untreated air containing carbon monoxide. The resulting
differential signal is exclusively due to carbon monoxide, since both cells
receive the same gas stream except that carbon monoxide has been removed prior
to entry in the reference cell. Because of the highway proximity, a 0-50 ppm
range was selected for the analyzer corresponding to 0-5 VDC output to the
computer. Tests performed at the site indicated a 60 sec. rise time to reach
100% concentration.
The NDIR analyzer also utilized an auto-zero/auto-span module during
operation. This is an electronic device which at pre-selected times auto-
matically gates zero gas to the analyzer and corrects the output from the
recorder to zero; then gates a calibration gas to drive the analyzer upscale,
correcting the output to the appropriate value. An out-of-range circuit is
also included in the system and this activates indicator lights when the zero
and span counters are at the limits of their count capability. The operator
is then required to attend to the instrument.
All calibration gases used were in the range of 40 ppm CO in hydrocarbon-
free air, doubly certified and purchased from Scott Environmental Technology
Inc. Four intermediate ranges of CO concentrations were also used to ensure
linear response of the NDIR analyzers.
11
-------
N
0 CO Prob«
A(2M)
€
p c
p €
p €
p €
^^i
p C
Y////A
p €
p ^
p
1 2 3 456 7 8 9 10 11 12 11
Trailtr
Fig. 4 Modified CO sampling Mt-up
-------
INFRARED SOURCE
Ul
o
UJ
o
UJ
UJ
u.
UJ
K
Ul
I
ift
olo
o_o
8?
3°«°0
UJ
I
CHOPPER^
0°«°^
0°"?0
O
UJ
IN
?-°c
OBO
°o.£
:>•$
!lc8b°o° SAMPLE OUT
DIAPHRAGM
DISTENDED^
DETECTOR
COMPONENT
OF INTEREST
o OTHER MOLECULES
Fig. 5 Functional Diagram of a NDIR
13
-------
NDIR MODIFICATIONS
During the zero mode, it was noticed that the output generally deviated
from the original setting. This deviation occurred even with replacement of
the detectors. The NDIRs, as originally configured, had a zero-span cycle
time which lasted 25 min. Further, the computer routine did not have the
capacity to record negative "zero" output from the NDIR. In order to account
for the zero drift, two modifications to the zero-span module's circuitry
were necessary.
The first alteration involved the inclusion of a potentiometer into the
zero-mode circuitry of each NDIR. These were mounted external to the NDIRs.
Each potentiometer was adjusted until a voltage of 250 mv was obtained across
the circuit, which allowed the zero-mode output of each NDIR to be set
initially to 250 mv. Thus, a downward decline, corresponding to -2.5 ppm, was
still sent out as a positive signal. The span adjustment was also set 250 mv
higher than before to obtain linearity throughout the range. Raising this
zero setting allowed the measurement of zero drift below the set zero without
increasing the computer memory storage space, which would have been required
for negative voltages. Most zero drifts were confined to + 0.5 ppm.
The second modification was also necessitated by the zero-span module
and the computer recording scheme. Both the zero and span cycles consisted
of two phases. During the first phase, the module monitors the current
uncorrected value of either zero or span. The circuitry adjusts for any drift
from the pre-set values during the second phase and outputs a corrected value.
These values are necessary for data reduction in order to establish the true
zero level for each of the detectors. An external switch which allowed the
selection of either a short or long time interval was attached to the cycle
time circuitry of the zero-span module. This facilitated manual adjustment
of the NDIRs when needed.
The short zero-span cycle lasted a combined length of 5 min. and was used
during correction of malfunctions of the analyzer. The longer zero-span cycle
lasted 50 min., and required the insertion of additional resistors to the time
constant circuitry. This position was the normal setting for the switch. By
increasing each (zero and span) cycle time to approximately 25 min., the
computer was able to record both the initial value and its final value,
regardless of when either cycle started. The long time was required because
of the averaging routine used for the vertical profile system. The design of
the vertical profile system is described in this section.
All NDIRs were automatically set to zero and span three times daily at
4 a.m., noon, and 8 p.m. These times were selected so that monitoring during
heavy traffic periods would not be interfered with. If a NDIR was not
operating properly, it was run twice through the short cycles before manual
adjustments were made. Normally the instrument corrected itself before manual
adjustments were necessary. After using the short cycle, the switch was
rethrown and the instrument was run through a long-cycle, so that the computer
could record the outputs.
14
-------
AVERAGING CHAMBERS
Rapid changes in CO concentrations were observed over short time periods
which were not compatible with the computer monitoring system. Although
these variations reflected the actual conditions, they caused inaccuracies
in the ten minute averages. This problem was apparent when comparing the four
reported values on the computer output for a common sampling point. The eight
instantaneous readings were insufficient to reflect the true integrated
average over the time period. To alleviate this problem, various size
averaging chambers were tried. It was found that 1 gal. bottles produced the
best results for the flowrate used. This volume was large enough to damp
out spikes, and yet small enough not to mask longer changes. Therefore,
averaging chambers were fabricated for all sampling points.
Averaging chambers were particularly useful for successful implementation
of the vertical profile system, since only two readings per sampling point
were taken in the 10 min. interval. The profile system incorporated sampling
points: 9A, 9B, 9C, and 9D (Fig. 6). Using a single NDIR, which eliminated
systematic differences between analyzers, consistent measurements of the small
difference in concentration between points 9A and 9B were obtained using the
averaging chambers.
Fig. 7 is a flow diagram of the profile system. The collector tank was
used for the profile system. Each sampling line led to an averaging chamber.
The destination of the flow from the exit port of each was controlled by a
three-way solenoid. During any given time, only one solenoid was activated.
Flow through this sample line was then drawn by the NDIR pump for analysis.
Flow for the other three lines was drawn by the main pump through the collector.
All flowmeters were set for a pressure corrected flowrate of 1.7 1/min. There-
fore, whether the sample was being analyzed or not, the flow through the
averaging chamber was maintained constant.
15
-------
N
D(16M)
C(8M)
B(4M)
A(2M)
Prob«
O
c
p c
p c
p c
p c
Y////A
p c
Y////^
p c
p c
p
1 2 3 456 7 8 » 10 11 12 13
Traitor
Fig. 6 CO Sampling system with the vertical profile system
-------
AVERAGING
CHAMBERS
9B
9A
3- WAY
SOLENOIDS
FLOWMETERS
WITH
NEEDLE VALVES
V
a
V
0
vv
0
D
PUMP
COLLECTOR
TANK
NDIR
NDIR
PUMP
Fig. 7 The flow diagram for the vertical
profile system.
17
-------
SECTION 4
PARTICULATE DATA COLLECTION
DESCRIPTION OF THE SAMPLER
«
Eight dichotomous samplers were custom-designed for this study by the
Environmental Research Corporation. The samplers had a nominal sampling rate
of 50 1/min and were capable of sampling from a 2 to 8m height. Fig. 8 shows
the location of samplers. The dichotomous samplers collect airborne parti-
culates in two size ranges (<3.3|j,m and >3.3um) on separate filters. Fig. 9
is a schematic of the sampler flow system. Particle separation is achieved by
means of a two-stage virtual impactor. The virtual impactor draws its sample
isokinetically from a 200 1/min flow stream from which particles greater than
18um have already been excluded. This removal is done via an auxiliary flow
(20 1/min) around the perimeter of the sampling head. Particles above the
cut-off point size (3.3um) flow directly through the central nozzle of both
stages of the impactor and are collected on the coarse particle filter. The
major portion of air sample is deflected around these nozzles and contains
the particles below the cut-off point which are subsequently deposited on the
fine particle filter. Flow tubes after the virtual impactor create a uniform
deposition across both filter faces. Flow through the sampler head and its
auxiliary flow is produced by a standard hi-vol blower. Differential pressure
gauges are used to monitor these flows. A retention filter covers the blower
discharge so that these particulates and the carbon worn from the brushes will
not bias the sample being collected. The top of the sampling head has a circu-
lar, lipped cover which prevents rain, snow, and road salt-sand from entering
the sampling train. Flows through the virtual impactor are drawn by a carbon
vane pump. These flows are monitored by rotameters, while needle valves are
used to set the proper flowrates. The system is designed for flowmeter
readings of 1.7 1/min for the coarse particle stream and 48.3 1/min for the
fine particle stream. The coarse particle flow operates near atmospheric
pressure, while the fine particle flow is at 7 psig vacuum. Therefore, the
actual flow corrected for pressure for the fine particle stream is 35.3 1/min.
The sampler head, virtual impactor, filter assembly and blower are
contained in a single unit which can be raised if a sampling height greater
than two meters is desired. A metal plate with a ring welded to the upper
surface was attached to the top of the sampling unit by three steel rods. The
plate allows the sampler to be raised, as well as providing additional pro-
tection to the sampling head from rain and snow, The virtual impactor pump,
flowmeters, gauges and elapsed time indicator are housed in a control box on
which the collection unit normally rests.
18
-------
N
/\ Dichotomoue Sampler
(5M)
(2M)
A
V////A V7///A
A
1 2 3 466 7 8 9 10 11 12 13
I Trailer
Fig. 8 Dichotomou* sampler location in the sampling plane
-------
AIR
RETENTION
FILTER
HI-VOL
BLOWER
150 LPM
VIRTUAL
IMPACTOR
48.3 LPM
.7 LPM
COARSE
FILTER
Fig. 9 Dichotomou* sampler flow pattern
20
-------
Millipore Fluoropore filters (0.5(jm pore size, 47mm diameter) were used
for sample collection. When enclosed in the sampling ring, the exposed
diameter is 40mm. This gives an effective deposition area of 12.6 cm^. An
0-ring imbedded in this filter holder prevents both sideway leakage through
the casing as well as cutting of the filter membrane.
The cut-off point of each sampler was determined by the manufacturer
before delivery of the equipment. These tests were performed for aerosols of
2, 3, 4, and Sum in diameter at a 50 1/min flowrate. The percent of aerosol
concentration on each filter and the overall losses for each of these size
particles was calculated. Fig. 10 is a composite diagram for all samplers.
For individual samplers, the 50% cut-off point for coarse particles varies
between 3.2um and 3.35um as determined by interpolation. As can be seen from
this figure, losses are of the order of 5% for aerosols between 2 and 5um
diameter.
For two samplers, the cut-off point was estimated for flows of 40 and
60 1/min as well. Fig. 11 shows these results. As can be seen, collection
efficiency decreases as the flow varies from the design rate. The cut-off
point also shifts inversely with a change from design flow.
PreHminary Tests
Before the samplers were deployed along the sampling plane, all of them
were stationed at Tower 11 to determine consistency of operation and compara-
bility between the samples from different units. Ten runs were made with all
samplers operating at approximately the same distance from the highway.
Sample periods of 60 to 150 min. were tried.
The first five runs were made with all settings initially at manufacturer
specifications. The fine particle air stream flow dropped 5 to 10% during the
sampling period, while the coarse particle air stream flow sometimes rose by
6%. The sampler head flow varied + 15%. These changes occur because the
samplers have no flow controller to compensate for filter loading, or changes
in ambient conditions.
Since the minimum flow rate through the sampling head is critical, the
decision was made to set all the samplers at an initially high setting to
ensure that the flow never went below the allowable limit. It was decided to
set the fine particle flow 3 percent higher than the design flow, so that it
would drift down through the design setting. Since the coarse particle flow
remained the same or increased only slightly, the setting for the coarse
particles was left at the design rate. Setting the fine particle flow
initially high allows the sampler to operate usually within + 3% of the design
rate for the entire sampling period, while the average flow remains within
+ 2% of the design flow. Referring back to Fig. 10, it is apparent that this
operation minimizes cut-off point shift and particle losses.
The second set of five runs was performed with the above changes in
initial flow settings. Flow readings were recorded every half hour during
these runs. Sample air volumes computed by taking only the end-points were
compared to the average volume computed from these and the intermediate values.
21
-------
100-
801
o
§60
oe
UJ
40
20
Fine
Concentration
Coarse
Concentration
12345
AEROSOL DIAMETER ( ji)
(Unit Density)
Fig. 10 % of aerosol concentration and overall losses
22
-------
100
D 40 L/MIN
O 50 L/MIN
A 60 L/MIN
Fig. 11
234
AEROSOL DIAMETER ( p )
(Unit Density)
% of aerosol concentration for different
flow rates.
23
-------
The largest difference between these two methods was 2%. Therefore, it was
decided that taking start-up and shut-off readings would be satisfactory.
All filters were returned to Albany for laboratory analysis. Comparison
of results between different sampling intervals led to the conclusion that a
two-hour sampling period yielded overall the most satisfactory results. This
was a compromise between the need to minimize the variation of traffic, wind,
and flow rates during sampling, scheduling of other duties at the site, and
collecting a satisfactory filter loading.
Iron pipe (2 1/2" OD) was used as booms for the three elevated sampling
points. A pulley was attached to the boom 1m from the tower. The samplers
were raised by means of nylon rope to a height of approximately 5m.
SAMPLING PROCEDURE
Table 3 is a copy of the log sheet used for sampling during the study.
On this sheet, "S" stands for the "small" (fine) particle filter, while "L"
denotes the "large" (coarse) particle filter. Each filter was pre-weighed in
Albany and placed in a self-locking plastic petri dish. The shallow height
of these dishes restricted the filters from curling and prevented the filter
surface from touching the dish surface. Each dish was alpha-numerically
labelled and sequentially packaged (e.g. 1A to 90A, IB to 90B, etc.). The
drawing in the upper right is a representation of the filter ring cassette.
Even numbers were used exclusively for the coarse particle filters, while odd
numbers were used for the fine particle filters. Filters were used sequen-
tially. This convention was adopted as an aid in keeping track of the filter
placements by the operator during sampling. Filters were mounted in and
demounted from the filter rings inside the trailer. All filter placements and
removals were done with the use of tweezers. If the filter was contaminated
for any reason during these procedures, it was discarded.
All filter cassettes were put in the samplers and the elevated samplers
raised into position before the first sampler was turned on. As each sampler
was started, the flows were set at the values indicated on the sheet. "L Mag"
refers to the differential pressure gauge on the left side on the control unit.
This gauge measures the flow through the sampler head. "R Mag" is the
differential pressure gauge on the right side, which measures the auxiliary
flow. These flows were controlled by a common voltage control to the blower.
While all the samplers were in operation, each was rechecked to ascertain
that all flows had their proper settings; if not, adjustments were made. Due
to the distance between samplers and the need to cross the highway, most runs
were done with an operator on each side. It usually took less than 5 min.
between start-up of the first sampler and the last. If only one operator was
present and traffic was heavy, this interval increased to about ten minutes.
During set-up, the date, weather condition, general wind direction, sampler
locations, and time of day of start-up of the first sampler were noted. The
elapsed time indicator was also reset to zero. At the conclusion of sampling
all final flow readings were recorded as well as the duration of sampling.
The samplers were shut off in the same order as they were started. The time
24
-------
TABLE 3. LOG SHEET USED FOR THE DICHOTOMOUS SAMPLER
DATE:
OPERATOR
- ^**.*_n^.i. WWJ-IA-* j. \_>±\ iiiu u -n_aj.\_» j.v/1'ivuo OriiiJrLiIjK.
WEATRER:
OR: TIME PERIOD: TO
®
EVEN
OHD
Filters
1
2
3
4
5
6
7
8
s ( )
1 ( )
Position ( )
Time f >
S ( )
L ( )
Position ( )
Tim? ( )
S ( )
L ( )
Position ( )
Time ( )
S ( )
t ( )
Position ( )
Time ( )
S ( )
L ( )
Position ( )
Time ( )
S ( )
L ( )
Position ( )
Time ( )
S ( )
I. ( )
Position ( )
Time ( )
S ( )
L ( )
Position ( )
Tine ( )
Wind Direction
Wind Speed
Gauges
L Mag (in. B20)
R Mag (in. H20)
Sm. Flow (SCFH)
Is. Flow (LPM)
L Mag (in. H20)
R M«g (in. H20)
Sm. Flow (SCFH)
Lg. Flov (LPM)
L Mag (in. H20)
R Mag (in. H20)
So. Flov (SCFH)
Lg. Flov (LPM)
L Mag (in. HjO)
R Mag (in. H20)
Sm. Flow (SCFH)
Lg. Flow (LPM)
L M»g (in. HjO)
R Mag (in. H20)
Sm. Flov (SCFH)
Lg. Flow (LPM)
L Mag (in. H20)
R Mag (in. H20)
Sm. Flow (SCFH)
Lg. now (LPM)
L Mag (in. H20)
R Mag (in. H20)
Sm. Flow (SCFH)
Lg. Flow (LPM)
L Mag. (in. H20)
R Mag (in. H20)
Sm. Flow (LPM)
1st. Flov (LPM)
2A
12A
ZA
12A
Initial
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
105 ( )
1.7 ( )
1.9 ( )
50 ( )
1.7 ( )
Final
Comnents
25
-------
of day when all sampling had ended was recorded. At this point, the elevated
samplers were lowered and all filters were collected and returned to the
trailer.
The sampling times were usually within the range of 120 + 10 min.
although intervals as low as 90 min. and as high as 150 min. occurred
occasionally due to the necessity of performing other activities at the site.
The difference in time between samplers in any particular run normally ranged
no more than 5 min. between the longest and shortest period.
Field Sampling
Cross-plane sampling was divided into two phases. The first phase con-
sisted of 71 runs, including the tracer study periods. All filters used
during this phase were pre-weighed in Albany and returned for subsequent
analyses. Generally, two runs per day were made. The first run in the
morning began around 9 o'clock, while the second was started in the early
afternoon about 2 o'clock.
The second phase of field sampling, consisting of 165 runs, differed
from the first in two major ways. Because of the low concentrations being
collected on the coarse particle filters, generally below the detection limit
of the x-ray fluorescence analyzer, these filters were no longer analyzed
after sampling. The second difference was the time scheduling of runs.
Arrival times at the site by the field personnel varied from 6:30 a.m. to
9:30 a.m. An early arrival allowed the first sampling period to end before
the morning rush hour traffic started to decline. Inversely, a late arrival
allowed the second sampling period to start after the onset of the afternoon
rush-hour traffic. Incidentally, the term "rush-hour" is a bit of a misnomer
when applied to the site since it generally lasted three hours.
Five runs during the second phase were made for special studies by the
New York State Department of Health. Sampling times of 30, 60, and 120 min.
were used during one day, preceded and followed by overnight runs of 1,000
min. Nuclepore filters with a 0.4um pore size were used during this period.
These filters are being analyzed for trace metals and particle size.
Sixteen additional runs were made at the conclusion of the study with
all the samplers again stationed at Tower 11. This was done to gather more
data to be used for error analysis, as well as to ascertain whether the
samplers had undergone any major changes in their sampling characteristics
due to maintenance.
26
-------
SECTION 5
TRAFFIC DATA COLLECTION
Vehicular counts by vehicle length and speed categories were obtained
from each lane of the highway by means of induction loops and ancillary
electronics. These were then stored in the computer and every 10 min an
output typed out for each direction, containing the total number of vehicles
according to speed/length category. Once a week, time-lapse movies of the
traffic were made to determine the age composition of the vehicles as well as
to provide a check on the electronic traffic monitor. Visual counts were
carried out by field personnel when deemed necessary.
A pair of induction loops were imbedded in each lane of the expressway.
The separation between the pair was 11 feet. A control box on each side of
the roadway was used to house the loop detectors (Streeter Amet Model 740
with solid state output). The signal from the loops was carried to the
trailer via multiple pair shielded cable. Line driver/receiver circuitry in
the trailer served as an interface with the computer.
The computer polled each loop for vehicle presence every 16.7 milli-
seconds. If the up-stream loop indicated a vehicle was present, the computer
measured the time interval until the second loop came on, as well as the
total time the first was on. Since the distance between the loops is known,
the vehicle speed can be calculated. Using this speed, the vehicle length
can be calculated. This speed-length measurement was then stored in a 10 x 5
array as shown in Table 4.
Problems were encountered with the sensitivity setting of the loop
detector. This was inferred when traffic counts printed out by the computer
were suspected to be low or when one of the length categories had too many or
too few vehicles. Then, adjustments of the detectors were made and visual
traffic coxrnts were undertaken to ensure that the corrections were in the
right direction.
A total of 32 time-lapse movies of the traffic were taken during the
study with a photographic system, manufactured by Time Lapse Incorporated.
The camera was set at a speed of 4 sec a frame. The camera operates via a
rechargeable battery with an internal clock and the film cassette used would
run for 4 hr uninterrupted. The camera was mounted on the overpass bridge of
Pinelawn Road (see Fig 1). Either direction of traffic flow could be photo-
graphed. Only days which were not cloudy, rainy, or foggy were useful for
photography. Before each run, the first few frames of the film was exposed
to a date and direction sheet; once the camera is started, with the internal
27
-------
clock set, it registers the time of the day on each of the frames. After the
run was over, the camera was brought back to the trailer for recharging the
battery and the film was sent out for professional development.
TABLE 4. LENGTH & SPEED CATEGORIES
Vehicle Lengths) (ft.) Speed (tnph)
< 11 Stalled
11 - 24 Stalled - 5
24-35 5-10
35 - 46 10 - 15
> 46 15 - 20
20 - 30
30 - 40
40 - 50
50 - 60
> 60
28
-------
SECTION 6
METEOROLOGICAL MEASUREMENTS
Meteorological data were collected from a total of 17 instruments.
These are:
5 CI-3 wind system
1 CI-25 wind system manufactured by Climet Instruments Co.
1 CI-18 wind system
4 Gill UVW's manufactured by R. M. Young
2 Temperature difference sensors
1 Temperature sensor
1 Humidity sensor manufactured by Meteorology Research,
1 Rain accumulator Incorporated
1 Pyranometer manufactured by Spectrosun
The wind instruments were mounted on the tower by means of a 2 1/2" x 5'
iron pipe clamped with U bolts. Attached to this was a 90° reducer to 1"
from which a 1" x 2' pipe extended vertically to support the equipment. From
each of the supporting beams a 2/0 copper wire was attached to a grounding
rod driven at the base of the tower for removal of static charges, accumulated
on the towers.
CLIMET INSTRUMENTS
The five CI-3 wind measuring systems with translators and Esterline-
Angus chart recorders were obtained on loan from USEPA in Research Triangle
Park, North Carolina. The CI-25 was purchased from the manufacturer to
complement the CI-3's. Both models were designed by the manufacturer so that
the anemometer and the vane are separated by a 3 1/2 ft bar. This bar was in
turn mounted on the vertical 1" pipe. A four-lead shielded cable connected
the wind speed measuring heads on the towers to the translators in the
trailer, while a three-lead shielded cable was used for the directional heads.
The principle of operation of the CI-3 and CI-25 is similar. The anemometer
which measures the wind speed consists of a cylindrical slotted drum which
alternately masks and exposes light from a lamp to a photo diode, producing
pulses at a frequency proportional to the rate of rotation of the propeller.
These signals are then sent to the translator for conversion to a B.C. analog.
Horizontal wind direction is measured by the 0-360° wind vane system. The
stem of the vane system is coupled to a 10K-100K high precision potentiometer,
29
-------
which produces an output corresponding to the attitude of the directional
vane. The wind vane systems were calibrated to output 0-5VDC for a direc-
tional range of 0-360°. However, it was noted that this 0-360° capability
of the CI-3's and CI-25 posed a problem of ambiguity and error when the vane
fluctuates between the fourth and first quadrant. The analog output will be
5V for 360° but 0V for <1°. On averaging this, the computer would yield 2.5V
(180°) or a direction directly opposite that from which the wind came. One
solution was to handle the problem via the computer program, but due to the
limited core capacity it was decided to purchase a Wind Direction Translator
from the Climatornics Corporation. Wind direction signals from the sensing
vanes of all CI-3's and the CI-25 were diverted to this translator, which
contained the logic for converting 0-360° transmission to a 0-540° signal at
a low impedance output of 0-5 VDC. Data integrity of the wind vane systems
was achieved by orienting the vane until a 0 VDC output was read. The vane
was then taped and mounted on the tower receptacle pointing to geographic
north. The orientation of the vane was done with a transit taking into
account the declination angle of 12°15" and the known orientation of the line
of sampling towers. Calibration of wind speed for both the models was done
such that 0-5 VDC corresponded to 0-50 mph.
Model CI-18 was on loan from SUNYA. Because of its age and prior use,
some of its components malfunctioned and it was returned to the factory for
repairs on two occasions. Even with much expenditure of resources and man-
time the instrument failed to perform satisfactorily. Eventually, all
attempts to obtain data from point 12A with the CI-18 were abandoned.
GILL UVW INSTRUMENTS
The Gill UVW anemometer is a three component wind instrument designed
for direct measurement of the 3 orthogonal wind components (east-west, "U";
north-south, "V"; and vertical, "W"). Three helicoid propeller sensors are
mounted at right angles to each other on a common mast, with sufficient
separation between propellers so that there is no significant effect of one
upon another for all normal wind measurements.
Four Gill UVW's were used at the project site. When in operation, a
miniature 2400 mv DC generator attached to each propeller shaft provides an
analog voltage output which is directly proportional to wind speed (a signal
of 5 VDC corresponds to 20 m/sec). A two conductor cable relayed data from
each sensor to a custom made UVW translator in the trailer. This unit was
capable of providing signal conditioning and filtering from the 4 UVW
anemometers. The output signals were then fed into the back panel of the
computer.
Bi-polarity is required since the anemometer measures forward and reverse
air flows; when the propeller rotation reverses, the polarity of the generator
also reverses. Response of the propeller as a function of its orientation to
the wind vector closely approximates the cosine law. However, since this
response is not exact, correction was made for all data points using an
algorithm developed by Horst (1971).
30
-------
Calibration of the UVW's was done by means of a unit with a calibrated
1800 rpm rotation connector shaft. The shaft was attached to the propeller
mounting rod by means of a short tubing, and rotated in each direction in
turn. At 1800 rpm's the trim potentiometer of the horizontal components were
adjusted to 2.25 VDC to correspond to a 9 m/s wind and the vertical system
set to 2.83 VDC for a 11.3 m/s wind.
OTHER INSTRUMENTATION
Temperature and Temperature Gradient
Temperature and temperature difference between two levels were measured
with Meteorology Research, Inc. Model 840-2 and 840-1 instruments. The
Model 840-2 placed at 4m on Tower 11 has two sensing elements, one for
temperature and the other the lower half of a AT circuit; Model 840-1
containing the second element for AT was placed at 15m.
Each temperature sensor is comprised of a dual thermistor and resistor
network, which provides a linear resistance change with an air temperature
change. Two dual thermistors are used in the AT sensor heads and each works
in a separate resistor network. The sensors are shielded by a power
aspirated, reflective cylindrical housing which provides a high heat transfer
from the ambient air to the sensing element, while at the same time affording
protection from incoming short-wave radiation and outgoing long-wave
radiation. The rate of air flow is approximately 15 f/sec.
Ranges for the sensors are -30° to +50°C for temperature and -3° to +3°C
for AT. Signals from both AT elements were transferred via cables to the
transmuter, where a circuit card converted the resistance difference to a
proportional DC output. The transmuter output for both T and AT was 0-5 VDC
corresponding to the ranges indicated above.
Relative Humidity
A Meteorology Research, Inc. Model 840-7 humidity sensor was placed on
top of the trailer for measurement of relative humidity. The element, placed
within a power aspirated, reflective cylindrical shield, is exposed to a
constant flow of air. The sensor is made from an assembly of organic and
inorganic crystals which detect moisture by the hydromechanical stress of
cellulose crystallite structures acting on a pair of thermally matched,
unbounded silicon strain gauges connected as half of a Wheatstone bridge.
The signal from the sensor is fed into a transmuter via a cable. The analog
output of the transmuter was connected to the computer back panel. Calibration
was done such that 0-5 VDC output from the transmuter corresponded to 0-100%
relative humidity.
Precipitation
A Meteorology Research, Inc. Model 302 rain gauge was placed on the roof
of the trailer for measurement of precipitation. The instrument employs a
7.86" diameter collector tube whose funnel is 8" below the upper rim for
31
-------
maximum collection efficiency even during strong winds with high rainfall
rates and a molded epoxy bucket pivoted on knife edges in Delrin wedges. A
water guide over the center line of the bucket assures equal fill and
eliminates splash loss.
In operation the rain gauge measures 7.95 cc of water when the bucket
over-balances and swings to the other side. A magnet mounted under the
bucket passes close to a magnetic switch during the tipping action causing a
momentary closure of the switch. This pulse is then relayed directly to an
MRI transmuter which, in turn, was connected to a recorder and the computer.
Each bucket tip of 7.95 cc of water corresponds to 1/100" rainfall. The
system was calibrated such that a 0-1" rainfall corresponded to 0-5 VDC out-
put from the transmuter. The calibration procedure consists of manually
tipping the bucket 100 times and accurately adjusting zero and span for
outputs.
Solar Radiation
A Spectrosun Model SR-71 pyranometer, on loan from the EPA, was mounted
on top of the trailer. Calibration of the instrument was achieved by compar-
ing its output to that of a recently factory calibrated standard, a Model
8-48 Eppley pyranometer with a Model 1040 Transmation digital potentiometer.
Both pyranometers were exposed to sunlight simultaneously and a digital
potentiometer was used to measure their output in millivolts. To ensure
proper calibration, several readings were taken during early afternoon. A
calibration correction factor K is obtained from the following equation:
K = mv (station sensor) X calibration of standard sensor mvc
mv (standard sensor) present calibration of station sensor mvs
The objective was to calibrate the system for a full scale range of 0-2
Langleys. The calibration against the factory calibrated standard sensor
showed that 2 Langleys corresponded to a sensor output of 10.16 millivolts.
The sensor signal was connected directly to a strip chart recorder calibrated
for a full scale sensitivity of 10.16 mv and a retransmitting slidewire
stepped up the voltage range to 0-5 VDC before transmitting it to the
computer.
32
-------
SECTION 7
TRACER GAS EXPERIMENTS
One of the important variables in the dispersion models is the source
strength. Since the gaseous and particulate emissions from the vehicles can-
not be estimated accurately, measurements of source strength were made using
sulfur hexafluoride, SFg, as a tracer gas on the expressway. Knowing the
source strength and concentrations of SFg and CO at various locations adjacent
to the highway and assuming that the dispersion of SFg is similar to that of
CO, one can estimate the source strength of CO. The experiments were conducted
during three week-long periods during October and November, 1976. Altogether
twenty-three runs were performed.
FIELD SET-UP
Sampling Train Set-up
The sampling train consisted of a battery operated pump connected to a
five layer snout type mylar bag via an acrylic rubber tubing and a short
length of tygon bubble tubing. The pumps, bags and the bubble tubing were
manufactured by Calibrated Instruments, Inc. The pumps, having a flow rate
of approximately 0.7 1/min, were mounted at the locations shown in Fig. 12.
The wiring of the pump was modified so that the batteries and switch required
for operation were external to the casing. Between the acrylic tubing coming
down from the pump and the sample bag, the bubble tubing was inserted firmly
and taped to ensure no leakage. This sample collection scheme is shown in
Fig. 13. The sample bags were then placed in heavy duty garbage bags which
were inside large plywood boxes. This was done to protect the sampling bags
from the wind and other factors that could cause damage. Once all the bags
were labelled and placed in position, the pumps could be turned ON/OFF by
switches located on the ground. The capacity of the bags was 44 1 which
corresponds to about an hour of sampling. Once the run was over, the bags
were separated from the bubble tubing and the snout was folded and taped. All
the bags were then collected and brought back to the trailer for analysis.
Tracer Gas Release Set-up
Six 1976 Plymouth Fury station wagons were used in this experiment. A
T-size cylinder containing 99.9% pure SF6 gas was placed on the floor of each
vehicle. The cylinder was blocked by wooden cradles with rubber trimmings so
that no lateral movements were possible when the vehicle was in motion or came
to an abrupt stop. A single stage regulator connected to the cylinder was
used for 'coarse setting1 of the SF£ release rate and for monitoring the tank
33
-------
N -
D(16M)
C(8M)
B(4M)
A(2M)
CD SFg Sampling Point
CD
l_
r H
_j i_
_i
i_
-i
L_
Y////A
_i
VS///A
L_
_)
l_
r S
_i i_
_i
1 2 S 456
0 9 10 11 12 13
Traitor
Fig. 12 Sampling locations for
-------
Acrylic Rubber
Short Length Tygon
Bubble Tube
Tube
Sample Inlet
Holes ~~
Air Pump /
0.7 lpm,1.5vdc
5 Layer Mylar Bag
44 Liter Capacity
Sealed Joint
Wrapped With
Duct Tape
Silicone Reseating Septum Disc
Sample Removal
Fig. 13 SF6 sample collection
35
-------
pressure. This, in turn, was connected to a pressure gauge and flowmeter,
which had been calibrated by Brook's Instruments, Inc., to operate at 75 PSIG.
These units allowed 'fine setting' of the SFg release rate. One end of a
1/4" diameter copper tube was connected to the flowmeter and the other end
was brought through the window of the vehicle to a point near the vehicle's
exhaust pipe. The tubing was firmly taped to the vehicle's body. This is
shown schematically in Fig. 14.
Procedure for Release of SFg
The route of the tracer release vehicles is shown in Fig. 15. Two
persons were assigned to each vehicle and the driver was instructed to drive
at 55 mph in the middle lane while the second person released the tracer gas
between the two bridges and recorded data in the log as shown in Table 5.
Prior to each run, the vehicles were lined up near the on-ramp at Exit 50
for leak checks and instructions on the flow rate setting. Once the cali-
brated pressure gauge and the flow meter readings were set at the prescribed
level (3.6 SLPM), valve #3 was closed, and the vehicles were ready for the
experiment. The vehicles departed at intervals of one-and-a-half minutes,
and the tracer gas release controlled by opening or closing valve #3 (see
Fig. 14). In the event the vehicles became bunched up due to heavy traffic,
the vehicles would wait at the on-ramp of Exit 50 and restart to maintain
the timed separation. This was done to ensure the creation of a constant
line source between the release and end points on the roadway. At the start
of the experiment, personnel were stationed on either side of the highway and
the median strip to turn on the switches of the sampling pumps simultaneously
when the second vehicle was abeam to the sampling plane on its return lap.
This was done to ensure adequate dispersion of the SF6 along the highway in
both directions. The sampling was continued for one hour at the end of which
one of the ground personnel signalled the shutting off of the pumps, and the
end of the run to the vehicle personnel.
SULFUR HEXAFLUORIDE ANALYSIS
The analysis was performed with a portable electron capture Gas Chromat-
ograph (A.I.D. Inc., model 511-06) in conjunction with a programmable
Supergrator II integrator (Columbia Scientific Industries) and an Esterline-
Angus millivolt chart recorder. The chromatograph was tested in the labora-
tory at Albany before it was taken down to the test site. The carrier gas
used was argon with 570 methane, with a flow rate of 30 ml/min which corresponds
to a head pressure of 13 PSIG on the G.C. The oven temperature was 55°C,
The G.C. was switched on for at least 12 hours before it was used so as to
establish thermal equilibrium between the oven, injection port and the
detector.
A stainless steel 1 ml sample loop was used for the sample injection
with the column size being 6' x 1/8" stainless steel molecular sieve 5A° and
45/60 mesh. The retention time for the sample was about 0.8 min. This was
confirmed by injecting ultra pure gas.
36
-------
1/4' Copper Line
•SF*
Calibrated
Pressure
Gauge No. 3
Release
Tank
Pressure"
Tank
Valve
75PSIG
99.99%
3-
Spherical Glass Float
Spherical Steel Float
Valve No. 3
\ Flowmeter
Calibrated at 75 PS IG
Adjusting Valve
Zero to Max. Flow
Fig. 14 Schematic of SF^ gas release set-up
37
-------
TLigtoO BAGATELLE RD.
-, Traffic
O tights
EXIT 50
OLD EAST NECK RD.
1
O.J
0.2
5 Miles
'Mites
\
RELEASE
GAS
1
J
RELEASE
GAS
SAMPLING TOWERS
I
Trailer
PINELAWN RD.
EXIT /
49N/
/A
/
I
$
t3
Si
6
>j
Fig. 15 Route for the tracer gas release
38
-------
SHEST # of
TABLE 5. RESEARCH ON AUTOMOBILE POLLUTANT DISPERSION
SF6 TRACER STUDY
DRIVER
OPERATOR_
TANK #
CAR #
DATE
TEMPERATURE MAX_
(CAR) MIN_
RUN #
FLOWMETER #
TIME
(LOCAL)
DIRECTION
E/W
LANE
DURATION OF
RELEASE
PRESSURE
FLOW METER
TOP
.
1
t
BOTTOM
REMARKS
39
-------
Three standards, 0.1, 1.0, and 10.0 ppb SFg in ultra pure nitrogen,
manufactured by Scott Environmental Laboratory, were used for calibration
purposes. The base line as well as any instrument drift were established
using samples of ultra pure nitrogen. The base line and calibration tests
were performed each day before the samples from the bags were injected. It
was found that the 10 ppb SFg standard when calibrated against 0.1 and 1.0 ppb
yielded a reading of 7 ppb. When 10 ppb samples were diluted to half by
volume with ultra pure N2 and tested against 0.1 and 1.0 ppb standards, it
was found that the linearity of the instrument extended to at least 5 ppb.
Before each run fresh air samples in the vicinity of the highway as well as
near the trailer were obtained with syringes and tested for the SFg concen-
tration in the ambient air. Normally the SFg concentration was found to be
zero. This was done to ensure that no leaks were present from the SFg tanks
stationed near the trailer, and that the SFg released in the earlier experi-
ment was completely dispersed from the vicinity of the roadway. This was
necessary because at least two, one-hour release experiments were conducted
per day and occasionally the sampled bags were purged in the vicinity of the
trailer. Samples were taken from the bag, after it was thoroughly shaken,
using syringes and injected into the G.C. The number of samples drawn to
establish the concentration of the bag varied from a minimum of 3 to a
maximum of 6. After each bag, either a standard was run or occasionally the
system was purged with ultra pure N£ and then the standards were run. The
sampler bags were reused by flushing them with air three times and then ultra
pure nitrogen. After the bags were flushed, samples taken from these bags
indicated no measurable SFg concentration and that the bags were flushed
properly.
40
-------
SECTION 8
SPECIAL MICROMETEOROLOGICAL EXPERIMENTS
INSTRUMENTATION
In order to assess the effects of traffic on the turbulent structure of
the atmospheric surface layer, time records of the turbulent wind components
and temperature as well as wind profile data were collected on both sides of
the Expressway.
Sonic Measurements
The three component wind fluctuations were measured by a sonic anemometer,
Model PAT 311, constructed by Kaijo Denki Ltd. The design is a solid state
pulse type with a 20cm path length. The instrument is based on the difference
of arrival times of pulses emitted along and in the opposite directions of the
wind. A detailed description of its development and usage can be found in the
Japan - U.S. Joint Study Group Report (1971). Fluctuations greater than 20 HZ
can be resolved with the help of the sonic anemometer. Before each run, the
instrument was zeroed electronically by placing it in a "windless" box.
Calibration was performed at least once every day using a Model 556 Tektronix
oscilloscope with a time delay. Corrections for an apparent overestimation
and zero drift were incorporated into the data reduction procedure.
The sonic head was mounted at a height of 3m on a sturdy boom extending
3m away from the tower to avoid any contamination due to sonic reflection. It
was oriented such that the mean wind was centered between the two horizontal
sonic paths. A total of 36 hourly runs were made during the fall of 1976 and
late spring of 1977 with the anemometer mounted on either Tower 2, 9, or 11
(see Fig. 16). The runs were made over a variety of wind directions, stability,
and traffic volumes and speed conditions. Each hourly run was then broken up
into 15 min. segments to be used for processing and the average results of
these four segments were used for analysis.
Temperature Fluctuation Measurements
Records of temperature fluctuations were obtained using a fast response
copper-constantan thermocouple whose signal was amplified by using a Honeywell
Accudata system with a gain of 5000. Fifteen runs were made and data was
recorded on a Model 5600 Honeywell analog recorder. A zero reference voltage
was recorded before each run to avoid playback problems.
41
-------
N
0(16M)
C(8M)
B(4M)
A(2M)
D
C] Sonic H«od
0 D
V////X V////A
4 5 •
• 9 10 11 1t 13
I I Traitor
Fig. 16 Location of th* Sonic toad
-------
Wind Profile Measurement
Mean wind profiles were measured using two Thornthwaite Associates Inc.
systems, equipped with digital readout. The matched cup anemometers were
mounted on booms extending 1m away from the tower to minimize tower influence.
A polaroid camera system was used to take pictures of the digital readout unit
at intervals of 15 min. This allowed for automatic acquisition of continuous
data for up to 12 hr. The new wind profile system was mounted on Tower 9 with
the cups positioned at 1.5, 2, 3, 5, 8, and llm levels, and the old system
was mounted on Tower 3 with the cups at 2, 3, 5, and 8m levels. The runs were
made in late spring of 1977 over a 14 day period of which only four days had
both systems working simultaneously. Even with these limitations, simul-
taneous profiles under rather variable conditions were acquired.
DATA PROCESSING
To minimize aliasing in the spectra, the time records were played back
through an active low pass filter. The cut-off frequency was chosen to be
10 Hertz to obtain maximum aliasing reduction for the sampling rate used, as
suggested by Kaimal, et. al. (1968). The data were then digitized using a
Hewlett-Packard Data Acquisition System at a rate of 20 samples/sec. The
digital tapes were then decoded and analyzed on the UNIVAC 1110 computer.
From these data, spectra and cospectra were computed using a fast Fourier
Transform Technique (Cooley and Tuckey, 1965).
The wind profile data were organized according to (a) wind direction with
respect to the highway, (b) traffic volume and speed, and (c) atmospheric
stability. Profiles were drawn by averaging over at least 7 consecutive 15
min. runs. In this averaging process, checks were made to ensure consistency
such that any changes at one level were accompanied by similar changes at all
other levels. For a given situation, profiles from different days or times
were also compared in order to check for the validity of the data. The
atmospheric stability was determined by computing a parameter defined as
where z is the height of observation, L is the Monin-Obukhov length, k Von-
Karman constant, g acceleration of gravity, T mean ambient temperature, U*
friction velocity and W^Tl heat flux. Further details on the data handling
and reduction can be found in Sedefian (1977).
43
-------
SECTION 9
COMPUTER DATA ACQUISITION & REDUCTION TECHNIQUES
COMPUTER SYSTEM
In this section details of the computer and its accessories necessary
to collect and store data are presented. Necessary operations were:
a. collection of carbon monoxide
data
b. collection of traffic data
c. collection of raw meteorological
data (also perform some inter-
mediate computations)
In order to carry out these tasks, the necessary hardware and software were
built around a Data General Corporation (D.G.C.) NOVA 2 minicomputer having
8-K words of memory. The entire configuration is shown schematically in
Fig. 17.
The accessory components are:
ASR 33 - Teletype
Dual Drive Diskette Unit from D.G.C.
Analog/Digital Interface from D.G.C.
Digital Interface from D.G.C.
and a Kiethley 10-channel Scanner.
Collection of CO Data
The output signals from the 8 NDIRs are fed into the back panel of the
hardware cabinet and then routed to the first 8 channels on the Kiethley
Scanner. The appropriate CO signal is selected by the scanner and passed
to channel #31 of the A/D interface via the Cable C2. A 74 sec. cycle is
used to sequence the 4-CO banks by which time all the 8 NDIR signals have
been processed. Hence, during the 10 min. data collection period, each bank
is cycled 2 times so that eight 74 sec. cycles occur, with an 8 sec. dead
span. During the 74 sec. cycle, the CO samples flow through the NDIRs, which
ensures purging the previously held samples. Between the 72nd and 74th sec.,
the NDIR is scanned by the Kiethley Scanner at 250 milliseconds interval and
the signal passes on to channel #31 of the A/D interface.
44
-------
Traffic^
Inputs
DATA GENERAL NOVA 2
Diskette Interface
Kiethtey
Scanner
Diskette Drives
NDIR
Inputs
To Solenoid
Banks
I 1 I
Meteorology
Inputs
Fig. 17 Configuration of th€ computer and
its accessories for data collection.
45
-------
Traffic Data
Each lane on the expressway had a pair of induction Loops imbedded in the
pavement. The computer monitors the 12 loops every 16.7 milliseconds. The
time interval between tripping loop A and loop B (which have a physical
separation of 11 ft. on the roadway) is used by the computer program to
determine the speed of the particular vehicle. The time interval for which
loop A remains tripped due to the presence of the vehicle is then used to
determine the length of the vehicle. Traffic signals are fed into the Hard-
ware Cabinet and routed to the Digital Interface by Cable C3. Each of the 12
loops corresponds to one bit in a 16 bit digital word, which is processed by
the Digital Interface. During the 10 min. sample period, the speeds and
lengths of all the vehicles are stored in the computer memory. At the end of
10 min., this information is reported as number of vehicles for 5 length and
10 speed categories for the west bound as well as east bound lanes.
M£teorological Data
The output voltages of the 11 wind, temperature and temperature differ-
ence, solar radiation, and humidity sensors are fed into the hardware cabinet
and routed to the A/D interface via the cable C2. The signals are collected
every 250 milliseconds and block-averaged to obtain an effective sampling
interval of 1 sec. The data from Gill anemometers were corrected for cosine
response (Horst, 1971). The data from the precipitation sensor are acquired
every 10 min.
COMPUTER PROGRAMS
The software as configured has three basic programs included in one master
program. These include:
1. Data Acquisition
Program
2. A/D Test Program
3. Debug 11 Program
Although the Data Acquisition Program is the prime program, the other two
programs are useful utilities for testing and/or debugging hardware and
software.
The Data Acquisition Program functions have been covered in earlier
paragraphs. The A/D Test Program can be used to determine individual sensor
data values. A variable number of data points (up to 200) at a variable time
interval (10 milliseconds to 30 seconds) between each data point facilitates
trouble-shooting of problems associated with particular groups of sensors.
All 32 carbon monoxide channels, 31 meteorology channels, and the traffic work
channel are accessible to this routine.
The Debug 11 Program has been imbedded into the program in order to
provide on-site changes in program instructions and program constants.
46
-------
Computer Output
Every 10 min. , the data are averaged, manipulated, and stored on the
floppy disk. The floppy disk has one recording surface divided into 77 tracks
and each track contains 8 sectors. Data are transferred to the floppy disk
from the computer memory in blocks of 256 16 bit words which corresponds to
one data block per sector. Table 6 lists the arrangement of variables within
a given block.
Since there are 144 10 min. periods in a day, a disk can accommodate
four days worth of data. Hence, it contains four "Contiguous files" labelled
DAY 1, DAT 2, etc. with each file being 144 blocks in length. The diskettes
were changed every four days and brought back to the Central Office for
further processing. A sample of the output for a 10 min. period is shown in
Fig. 18 where the signal range 0-5V corresponds to 0-1024 A/D units.
Data Processing
The next stage of data processing is to read each block of data, and
perform the following tasks:
a. Except for the traffic data, all the other
variables are to be converted to their
standard units,
b. Compute the necessary meteorological parameters.
c. Transferr the data from the mini-computer to
Univac 1110.
The necessary hardware and software required are:
i. A NOVA 2/10 mini-computer with 24K word memory.
ii. A D.G.C. asynchronous line adapter module.
iii. An ASR-33 teletype
iv. A dual drive and a single diskette drive.
The entire configuration is shown schematically in Fig. 19. The 10 min. data
are then transferred to Univac 1110 using the asynchronous line adapter module
along with ancillary software and stored on a magnetic tape. Hourly averages
are then obtained and stored on a magnetic tape.
47
-------
TABLE 6. ARRANGEMENT OF DATA WITHIN A GIVEN BLOCK
Data
Year
Day
Hour
Minut e
CO (32 channels)
Traffic (Lanes 1-3) West bound
Traffic (Lanes 4-6) East bound
Climet Wind Direction
Gill u
Gill v
Gill w
Temperature
Difference Temperature
Climet Wind Speeds
Humidity
Radiation
Vertical component of CI-25
Ground (Kiethley)
Precipitation
Blank
U
SQRT (ZUH*wi/600)
SQRT (Eu*wi/600)
SQRT (ZVl*wi/600)
SQRT (ZW1*T/600)
SQRT (EU12/600)
SQRT (EV12/600)
SQRT (IW12/600)
SQRT (£U52/600)
SQRT (EV52/600)
SQRT (£W52/600)
SQRT (ZU102/600)
SQRT (£V102/600)
SQRT (XW102/600)
SQRT (ZU252/600)
SQRT (ZV252/600)
SQRT (IW252/600)
Blanks
Word(s)
1
2
3
4
5-36
37-86
87-136
137-143
144-147
148-151
152-155
156
157
158-164
165
166
167
168
169
170-177
178-181
182-185
186-189
190-193
194-197
198-201
202-205
206-209
210-213
214-217
218-221
222-225
226-229
230-233
234-237
238-241
242-245
246-256
48
-------
1977
7
9
20
YEAR
J
CO VALUES
22
39
36
63
69
42
49
45
36
94
36
29
WEST BOUND
0
0
0
0
0
0
0
0
0
0
687
20
25
67
2
371
70
1
76
7
10
4
24
32
u,v,w
69
69
69
0
3
0
0
0
TOTAL
0
3
0
0
0
TOTAL
699
49
1
75
-1
462
89
0
75
0
0
0
19
0
SIGMAS
74
75
75
0
1
0
0
0
WEST
0
0
0
0
0
EAST
750
18
22
64
-2
1013
78
1
72
5
2
4
26
26
FOR
68
68
68
86
91
58
29
58
107
27
32
51
66
41
27
68
37
73
30
97
96
41
41
TRAFFIC
0
0
0
0
0
BOUND
0
0
0
0
0
BOUND
825
79
4
61
-1
2
84
4
63
2
0
2
23
7
1,5,10
66
67
67
0
0
0
0
0
=
0
0
0
0
0
=
1024
-91
582
25
0
1
1
0
3
0
0
0
658
0
1
0
0
0
265
619
0
-8
SEC
0
0
0
1
325
15
4
22
0
92
4
2
0
598
59
-9
0
1
2
0
239
18
0
7
0
93
7
4
9
Tl
0(
ot
0
0
1
HOUR
MIN
0
18
2
0
0
0
42
6
0
1
0
0
0
0
0
0
0
0
0
1
WIND DIR.
WIND SPEED
GILL U
GILL V
GILL W
TEMP DELTA TEM REL. HUM
SOL.RAD GRND PRECIP
GILL THETA
GILL PHI
UH
UHW1
SUHWl
U1W1
V1W1
WIT
31 1 28 9
30 0 27 8
28 1 26 8
Figure 18. Typical Computer Output
49
-------
DISKETTES
24-K WORD
MINI - COMPUTER
ASYNC
LINE
UNIVAC
1110
AS3-33
TELETYPE
Fig. 19 Configuration of the computer and its accessories
for data analysis.
50
-------
SECTION 10
LABORATORY ANALYSIS
In order to determine the participate weight of the samples collected by
the dichotomous samplers, the millipore filters were weighed before and after
the sampling. Techniques for their weighing as well as the analysis of these
filters for sulfur and lead by x-ray fluorescence and for various anions are
indicated below.
PARTICULATE WEIGHTS
The filter membranes which were used in this study were teflon (Millipore
Fluoropore) with a pore size of 0.5 urn. The filter samples were weighed on a
Mettler M-5 microbalance which had been enclosed in a glove-box for the pur-
pose of maintaining constant humidity and temperature. Control of humidity
was obtained through the use of saturated salt solutions which were exposed
to the circulating air within the glove box. This relatively simple technique
was adapted from the procedures described in ASTM, E 104-51 for maintaining
constant humidity through the use of aqueous solutions. Calcium nitrate was
used in this application and produced a relative humidity of 5170 at 25°C.
Weighing accuracies of + 15 jag were obtained when a 10 millicurie krypton-85
beta source was positioned within the balance to neutralize any electrostatic
charges on the filter membranes.
X-RAY FLUORESCENCE SPECTROMETRY
After the particulate weights had been determined, the filter membranes
were analyzed for trace metals on a Siemens Model VRS wavelength-dispersive
x-ray fluorescence (xrf) spectrometer. The Siemens xrf spectrometer has been
modified to accept a ten position sample changer, and the entire system is
interfaced to a Hewlett-Packard Model 9810 programmable calculator for the
storing and processing of data on magnetic tape. Final data processing is
performed on a Hewlett-Packard Model 9830 programmable calculator with a nine
inch thermal printer. A molybdenum target tube powered by a K-4 generator
was employed throughout this study;, each particulate sample was analyzed for
lead and total sulfur by use of lithium fluoride (200) and graphite crystals,
respectively.
Because of the particular geometry employed by Siemens in the construc-
tion of their ten position sample changer, a significant amount of backscatter
from the aluminum sample cup is able to reach the detector and cause a
deterioration in the detection limits of the lighter elements, especially
aluminum. Therefore, it became necessary to design and construct from
51
-------
Plexiglas our own sample cups and mounting rings in order to present the
samples to the spectrometer for analysis. This new design effectively
reduced the amount of backscatter, improved the detection limits of the
lighter elements such as sulfur, and permitted the analysis for aluminum in
samples from other projects.
The calibration standards for this phase of the analysis were prepared
according to the following procedure. The appropriate metal aerosol was
produced by pumping an aqueous solution of an appropriate soluble metal salt
into a May Spinning Disc Aerosol Generator. The resulting metal aerosol is
subsequently deposited onto a 10 cm teflon filter membrane by use of a
vacuum pump in much the same manner that the actual particulate samples were
collected along the roadside. From the particular operating parameters which
were used for the aerosol generator, the size of the resultant particles fell
in the range 0-10 (am, with ninety percent of the particles less than five
microns. Since the aerosol production was constant with time, the amount of
particulates deposited on the filter membrane was linear with time. It was
convenient to prepare samples with five different particulate loadings to
cover the entire range of interest within this investigation. Three smaller
discs (37 or 42 mm diameter) could be punched easily from each of the ten
centimeter filters. Then the fifteen calibration samples were processed by
xrf spectrometer and the net pulse counts were used to judge the uniformity
of the deposit for each set of three discs as well as the suitability of each
set in covering the appropriate concentration range. After this step was
completed and all of the calibration samples were judged to be satisfactory,
one particulate filter from each set of three was destructively analyzed for
lead by atomic absorption methods while sulfur, in the form of sulfate, was
determined by ion chromatography. Data from the resultant analyses were
plotted against the initial xrf net pulse count to yield the required cali-
bration curves. The curves for both lead and sulfur exhibited straight lines
(correlation coefficients = 0.99) and the lower limits of detection were 170
ngm/cm^ and 21 ngm/cm^, respectively.
ANION ANALYSIS BY ION CHROMATOGRAPHY
After the completion of analysis by xrf techniques, particulate filters
collected in the field were analyzed for anions of interest by ion chromato-
graphy. There are various methods currently available for analysis of sulfate
and nitrate in ambient aerosols. Mulik et al (1976) showed the successful
application of ion chromatograph to the analysis of aqueous sulfate and
nitrate in ambient aerosols.
The principle involves separating the species of interest on an ion
exchange separating column, followed by removal of the background ions in the
eluent with a suppressor column leaving the sample ions undisturbed; they are
monitored by a conductivity cell connected to a meter and recorder. For the
analysis of sulfate and nitrate, the separation column contains a strong basic
resin and the suppressor column contains a strong acid resin. A schematic of
the flow system is shown in Fig. 20. The flow system consists of a separator
or analytical column, suppressor column, four solvent reservoirs, injection
valve with sample loop, two Milton Roy fluid pumps, conductivity meter and a
52
-------
I I
REGENERANT
I
. -REGENERANT
J * VALVE
MANIFOLD
WASTE
SEPARATOR
VALVE
ANALYSIS
REGENERATION—
ALTERNATE
Fig. 20 Ion chromatograph flow system
-------
valving system to direct the flow through various parts of the instrument.
The system uses air activated teflon slider valves throughout.
The samples were extracted overnight on a laboratory shaker in 6 ml of a
solution of 0.0030 M sodium bicarbonate and 0.0024 M sodium carbonate. The
samples were then loaded into a sample tray for introduction into a Dionex
Model 10 ion chromatograph for the general determination of chloride,
fluoride, phosphate, bromide, nitrate, and sulfate anions. This equipment
had been automated through use of a Technicon pump and Sampler IV and a
Valco high pressure injection valve. The synchronization of the various
pieces of equipment was accomplished through the use of a Columbia Scientific
Industry Supergrator III integrator with programmable external contact
switches. Each chromatogram was traced on a Hewlett-Packard recorder while
the integrator printed all data in parts per million.
/ >
A standard solution of aft five anions was made at 50 ppm in the carbon-
ate/bicarbonate buffer solution. Portions of this solution were then diluted
volumetrically to give solutions containing 25, 10, 5, and 1 ppm of all five
anions. These solutions were processed through the ion chromatograph and the
resultant data indicated that a linear response was obtained for chloride,
fluoride, phosphate, bromide, nitrate, and sulfate over the concentration
range 0-50 ppm. For the routine processing of normal samples, one of the
above standard solutions was injected three times and the final data were
averaged to obtain appropriate calibration constants before the unknown
samples were injected. It was also routine to end the analysis late in the
day with a single sample of the same standard solution. Review of this data
with that from the morning would enable the operator to determine the amount
of instrument drift through the day as well as the occurrence of an equipment
malfunction. The precision of this method was better than 2% for a 500 ml
sample loop.
It should be mentioned here that another analytical method has been in
wide-spread use previously for the analysis of sulfates, especially for
automotive exhaust applications. The specific method was referred to as the
barium chloranilate procedure. Originally, this method was to be used in
this study. However, from our own tests and from the General Motors
Experiment fCadle et al, 1976), the barium chloranilate procedure was found
to suffer from many interfering ions, e.g., chloride, nitrate, bromide,
phosphate, and carbonate. Since this project was a field study and sulfate
levels were anticipated to be fairly low, low levels of interfering ions
were also to be expected during analysis but could not be tolerated. There-
fore, the relatively old technique of ion chromatography which employs the
fairly recently developed ion chromatograph was chosen as the analytical
method for this particular effort. For the specific analysis of sulfate in
this study, there were no interferences.
54
-------
REFERENCES
Cadle, S.H_, D.P0 Chock, J.M. Heuss, and P.R. Monson; "Results of the General
Motors Sulfate Dispersion Experiment", Research publication, GMR-2107, General
Motor Research Laboratories, Warren, MI, 1976
Horst, T.W.; "A Computer Algorithm for Correcting Non Cosine Response in the
Gill Anemometer", Battelle, Pacific Northwest Laboratories, BNWL-1651-1,
Richland, WA, 1971, (pp 183-186)
Japan - United States Joint Study Group Report; "Development of Sonic
Anemometer and its Application to the Study of Atmospheric Surface Layers",
Disaster Prevention Research Institute, Kyoto, Japan, 1971
Kaimal, J.C., J.C0 Wyngaard, and D.A. Haugen; "Derived Power Spectra from a
Three-Component Sonic Anemometer", J. Appl. Meteor., 7, 1968, (pp 827-836)
Mulik, J., R. Puckett, B. Williams, and E. Sawicki; "Ion Chromatographic
Analysis of Sulphate and Nitrate in Ambient Aerosols", Analytical Letters,
9, 1976, (pp 653-663)
Sedefian, L.; "Some Characteristics of Turbulence Adjacent to a Major
Highway", M.S. Thesis, State University at Albany, 1977, (pp 44)
55
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing]
REPORT \O.
EPA-600/4-78-037
4. TITLE AND SUBTITLE
DISPERSION OF POLLUTANTS NEAR HIGHWAYS
Experimental Design and Data Acquisition Procedures
5. REPORT DATE
June 1978
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSION-NO.
AUTHQR(S)
S. Trivikrama Rao, Marsden Chen, Michael Keenan, Gopal
Sistla, Ramam Peddada, Gregory Wotzak and Nicholas Kolak
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
New York State Department of
Environmental Conservation
Albany, New York 12233
10. PROGRAM ELEMENT NO.
1AA601 CA-05 (FY-77)
11. CONTRACT/GRANT NO.
R-803881-01
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. NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Interim 9/75 - 3/77
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The major emphasis of this investigation centered on the collection of
particulate and gaseous pollutant data, and detailed micrometeorological data
in a non-urban setting adjacent to the heavily travelled Long Island Expressway.
The purposes for collecting the data were to (i) document the distribution of
sulfate, lead, total particuIfctes and carbon monoxide at an array of sampling
points adjacent to the highway; (ii) study tne micrometeorology associated with
the highway, with special attention to those parameters important in the
determination of atmospheric dispersion, (Hi) reevaluate highway air pollutant
emission factors from data gathered in tracer gas experiments; and (iv) examine
the applicability of existing highway air pollutant dispersion models. The
location of the sites and the experimental setup for collecting pollutant data
are described, and details of the data acquistion procedures are presented.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
*Air pollution
*Atmospheric diffusion
*Limited access highways
*Micrometeorology
*Exper1mental design
*Data acquisition
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
13B
04A
04B
148
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
_66_
22. PRICE
EPA Form 2220-1 (9-73)
56
------- |