MANHATTAN COMMUNITY BASED PARTICIPATE STUDY
FINAL REPORT
EPA Contract 68D30029
Work Assignment 3-115
Prepared for:
Mustafa A. Mustafa
U.S. Environmental Protection Agency, Region II
Air Programs Branch
290 Broadway, 25th Floor
New York, NY 10007-1866
Prepared by:
TRC Environmental Corporation
6340 Quadrangle Drive, Suite 200
Chapel Hill, NC 27514
Januarv 15. 1997
-------
MANHATTAN COMMUNITY BASED PARTICULATE STUDY
FINAL REPORT
EPA Contract 68D30029
Work Assignment 3-115
Prepared for:
Mustafa A. Mustafa
U.S. Environmental Protection Agency, Region II
Air Programs Branch
290 Broadway, 25th Floor
New York, NY 10007-1866
-------
TABLE OF CONTENTS
Section Page
1.0 INTRODUCTION » 1
2.0 PM STUDY 2
2.1 Study Design 2
2.1.1. Phase 1 2
2.1.2 Phase 2 6
2.2 Sampling and Analysis 8
2.2.1 Samplers 8
2.2.2 Sampling Procedures 10
2.2.3 Analysis Procedures 12
2.3 Operational Problems 14
2.4 Quality Assurance/Quality Control 15
3.0 STUDY DATA 17
3.1 PM10 Data 17
3.2 PM2.5 Data 24
3.3 Carbon Analysis Data 30
3.4 Data Quality Indicators 30
3.4.1 Overall Precision 30
3.4.2 Analytical Precision 34
3.4.3 Relative Accuracy 34
3.5 Other Data 37
4.0 DATA ANALYSIS 39
4.1 PM10 Summary Statistics for Each Phase by Site 39
4.1.1 Phase 1 39
4.1.2 Phase 2 41
4.2 Comparison of PM10 Concentrations Among Sites 43
4.3 Comparison of Phase 1 and Phase 2 PM10 Levels 46
4.4 PM2.5 Concentrations and the Proposed PM2.5 Standards 46
4.5 Carbon Analysis 48
, 4.6 Unusual Events and Weather 48
4.7 QA/QC Results .'. 49
4.7.1 Completeness 49
4.7.2 Precision 49
4.7.3 Relative Accuracy 49
5.0 CONCLUSIONS AND RECOMMENDATIONS 51
5.1 PM10 Results and Implications 51
5.2 PM2.5 Results and Implications 52
5.3 Performance of Saturation Samplers 53
CH-97-02 ill
-------
TABLE OF CONTENTS (continued)
Section Page
6.0 ACKNOWLEDGMENTS 54
7.0 REFERENCES '. 55
APPENDIX A DAILY PM10 AND PM2.5 CONCENTRATIONS BY SITE FOR PHASE 1
APPENDIX B DAILY PM10 AND PM2.5 CONCENTRATIONS BY SITE FOR PHASE 2
APPENDIX C COMPARISON OF PM10 AND PM2.5 CONCENTRATIONS MEASURED BY
REFERENCE AND SATURATION SAMPLERS
CH-97-02 IV
-------
LIST OF FIGURES
Number , Page
1 Locations of Reference Samplers and Areas for Phases 1 and 2 3
2 Site Locations for Phase 1 4
3 Site Locations for Phase 2 7
4 Diagram of Mounted Saturation Sampler 9
5 Sampler Installations at (a) Site 1 and (b) Site 5 for Phase 1 11
6 Mean PM10 Concentrations by Site for Phase 1 20
7 Mean PM10 Concentrations by Site for Phase 2 21
8 Daily PM10 and PM2.S Concentrations at Phase 1 Site 1 27
9 Daily PM10 and PM2.5 Concentrations at Phase 1 Site 9 28
10 Daily PM10 and PM2.5 Concentrations at Phase 2 Site 9 29
11 Average PM10 for Phase 1 Harlem Community by Site Type 42
12 Average PM10 Phase 2 Washington Heights Community by Site Type 45
CH-97-02
-------
LIST OF TABLES
Number , Page
\
1 PM10 Data (in pg/m3) for Phase 1 - Harlem Community 18
2 PM10 Data (in pg/m3) for-Phase 2 - Washington Heights Community 19
3 Phase 1 Harlem Community Site Rankings 22
4 Phase 2 Washington Heights Community Site Rankings 23
5 PM2.5 Data (in pg/m3) for Phase 1 - Harlem Community 25
6 PM2.5 Data (in pg/m3) for Phase 2 - Washington Heights Community 26
7 PM10 Carbon Analysis Data (in pg/m3) for Phase 2 31
8 Phase 1 Overall Precision Data from Collocated Samplers 32
9 Phase 2 Overall Precision Data from Collocated Samplers 33
10 Relative Accuracy Results for Phase 1 35
11 Relative Accuracy Results for Phase 2 36
12 Average Traffic Counts for Select Harlem Community Sites 38
13 Variability in PM10 Data for Phase 1 - Harlem Community ... '. 40
14 Variability in PM10 Data for Phase 2 - Washington Heights Community 44
15 Significance of Site Concentration Differences for Phases 1 and 2 ,47
CH-97-02 VI
-------
1.0 INTRODUCTION
i
The U.S. EPA Region D conducted this study to address the concerns of the northern
Manhattan communities of Harlem and Washington Heights regarding air paniculate levels in
their neighborhoods. The EPA Region n Office received numerous comments about paniculate
matter pollution levels in northern Manhattan. Community representatives expressed concern that
northern Manhattan communities have higher levels of asthma than other parts of Manhattan and
felt that diesel bus and truck traffic in their neighborhoods might be responsible.
The purpose of this study was to collect air quality samples of paniculate matter
10 microns or less in diameter (PM10) at numerous street level locations in two communities in
northern Manhattan, Harlem and Washington Heights, and at the two permanent PM10
monitoring sites, Madison Avenue in midtown and Canal Street in downtown Manhattan. In
addition, samples of paniculate matter 2.5 microns or less in diameter (PM2.5) were also
obtained at a subset of the monitoring sites to provide information on PM2.5 levels particularly
in anticipation of EPA's proposed national ambient air quality standards (NAAQS) based on
PM2.5.
The study of paniculate matter pollution levels in northern Manhattan was designed to
provide preliminary answers to two questions:
(1) Are PM10 concentrations in northern Manhattan comparable to, less than, or
greater than PM10 concentrations at reference PM10 monitoring sites in midtown and
downtown Manhattan?
(2) How much variation is there in PM10 levels among community locations and, to
a limited degree, to what extent do automobiles, trucks, and buses seem to contribute to
PM10 concentrations?
Results will also be used to assist the State of New York in determining whether
additional monitoring is warranted in these neighborhoods and, if so, where additional monitors
should be located.
Representatives of the affected communities were active participants in the study from the
initial meeting in June, where the project was presented to the communities and a preliminary
list of sites for both Harlem and Washington Heights was,selected, to the presentation of the final
report, which is slated for January 1997. In a series of eight meetings to date, community
representatives participated in the training given to field operators, took pan in tours of potential
sites for the communities, helped to select the final sites chosen, and have received preliminary
data as they became available. The sites chosen for placement of samplers included locations
where population exposure was likely due to vehicular emissions, where high volumes of diesel
truck and bus traffic were suspected, where roadway pollutants could be trapped, and near
populations with sensitive health concerns that were in proximity to high traffic locations.
Undergraduate and graduate students at The City College of New York, located in the Harlem
community, were used as field operators.
CH-97-02
-------
2.0 PM STUDY
2.1 Study Design
Air quality samples of PM10 were collected in two communities in northern Manhattan
over a period of approximately 80 days during July through September 1996. Sampling was
divided into two major phases, Phase 1, which focused on the neighborhoods of Harlem, and
Phase 2, which focused on the neighborhoods of Washington Heights. In addition, a one-week
pilot study was conducted prior to Phase 1 to try to identify and correct any equipment or
logistical problems prior to starting Phase 1. For the pilot study (Phase 0), daily sampling was
conducted at two sites, using two samplers per site to permit operational flexibility in servicing
the sites.
During Phases 1 and 2, New York State operated its PM10 reference samplers (Sierra
Andersen dichotomous samplers) on a more frequent schedule, approximately once every three
days instead of once every six days. Under this schedule the reference samplers always sampled
on the day required for EPA's one in six day sampling schedule. The extra sampling day was
, usually on the middle day (third day from the standard day) but would sometimes be shifted to
the second or fourth day so that the site would not have to be serviced on a weekend. As used
here, the term reference sampler means a sampler that has been designated by EPA as a Federal
' Reference Method or Equivalent Method for PM10 and thus has met specific design and
performance specifications.
Figure 1 shows a map of Manhattan and indicates the location of the reference sites in
downtown and midtown Manhattan and the general areas monitored in Phases 1 and 2. The
paniculate matter samplers used were saturation samplers supplied from EPA's Saturation
Monitor Repository (SMR) and are described in greater detail in Section 2.2.1. Sampling was
conducted midnight to midnight.
2.1.1 Phase 1
In Phase 1 of the study, which ran from July 10, 1996 through August 20, 1996, daily
PM10 sampling was conducted at eight sites in the more southern neighborhoods of northern
Manhattan, focusing on the neighborhoods in the Harlem/West Harlem/East Harlem communities.
Sites were chosen with community involvement to represent different neighborhoods within the
community and to reflect different types and levels of exposure. Figure 2 shows the locations
of the eight Phase 1 community sites. Descriptions and purposes of the community and reference
sites for Phase 1 are given below.
West Harlem Sites
I. 133rd Street between 12th Avenue and Broadway (bus depot and school) - On the south
side of the street, on lamppost closer to 12th Avenue (PM10, PM2.5)
2. 145th Street and Broadway (truck traffic and high population density) - On the west side
of Broadway between 145th and 146th Streets, center of the block (PM10)'
CH-97-02 2
-------
Figure 1. Locations of Reference Samplers and Areas for Phases 1 and 2
CH-97-02 3
-------
CH-97-02
Figure 2. Site Locations for Phase 1
4
-------
3. 125th Street and Amsterdam Avenue (high population density, shops, and bus and truck
traffic) - On the northeast corner traffic light (PM10, PM10 collocated sampling** on
reference sampler schedule)
Central Harlem Sites
4. Lennox Avenue between 147th and 148th Streets (Esplanade Gardens apartment buildings
and bus depot) - On the northeast corner of 147th Street and Lennox Avenue (PM10)
5. 135th Street and Lennox Avenue (Harlem Hospital and two public schools) - On the
southwest corner (PM10, PM2.5)
6. Edgecombe Avenue between 139th and 140th Streets (control site - little bus and truck
traffic). - West side of Avenue, center of block (PM10)
East Harlem Sites
7. Lexington Avenue between 99th - 100th Streets (bus depot) - On the northeast corner of
100 Street and Lexington Avenue (PM10)
8. 116th Street and Lexington Avenue (high population density) - On the northwest corner
of 116th Street and Lexington Avenue (PM10)
Reference Sampler Sites
9. Madison Avenue between 47th - 48th Streets (PM10, PM2.5, PM10 collocated sampling
on reference sampler schedule)
10. Canal Street at Broadway (PM10, PM10 collocated sampling on reference sampler
schedule)
* The sampler at this site was stolen on the first day of sampling. From July 13-25, 1996,
replacement samplers were located one block east at the northwest corner of Amsterdam
Avenue and 145th Street. The samplers were moved back to the original location on July
26, 1996 for the remainder of Phase 1.
** In collocated sampling, two samplers are located at a site within a few meters of one
another. The duplicate measurements are used to assess the degree of mutual agreement
of measurements (i.e., determine precision).
Daily sampling of PM2.5 was conducted at Sites 1 and 5 in Harlem and, according to the
State's sampling schedule, at reference Site 9. The PM2.5 sampling was performed as an
ancillary effort by EPA to obtain information on PM2.5 concentrations and their relationship to
PM10 levels in preparation for a proposed NAAQS for PM2.5.
CH-97-02
-------
During Phase 1 the PM10 reference samplers sampled on the following dates:
, July 10, 14, 17, 20, 23, 26, 30
August 1, 4, 7, 10, 13, 15, 19
Originally, daily sampling was to be conducted at the community sites for 28 days or until
the following data capture objectives were met: IS valid data points for each site plus a minimum
of 12 run days when the monitors at each location are collectively operational. However,
operational and logistical problems, described in greater detail in Section 2.3, caused data to be
lost and therefore the study had to be extended to 41 days in order to get enough samples to do
a comparison. At a few sites, not enough data were collected.
2.1.2 Phase 2
In Phase 2 of the study, which ran from August 22, 1996 through September 30, 1996,
daily sampling was initially conducted at seven sites in the more northern neighborhoods of
northern Manhattan, focusing on the Washington Heights community. In addition, daily sampling
at one Phase 1 site continued during this period in order to provide a data comparison between
* the two communities and to help account for variations in paniculate matter concentrations
' between the two phases. Monitoring on the enhanced reference sampler schedule of
'approximately once every three days was also continued at the two PM10 reference sites.
* Figure 3 shows the locations of the eight Phase 2 community sites. The reference sites for Phase
2 were the same as those for Phase 1; the descriptions and types of sites for the community sites
for Phase 2 are given below.
Washington Heights Sites
1. 181st Street and St. Nicholas Avenue (bus and truck traffic) - On northwest corner by the
Chemical Bank (PM10, PM2.5)
2. 179th Street and Broadway (bus depot) - On northwest corner (PM10, PM2.5)
3. 168th Street and St. Nicholas Avenue (hospital, school and bus traffic) - On northeast
corner by the school (PM10)
4. 162nd Street and Edgecombe Avenue (control site - little bus and truck traffic) - On
northeast comer (PM10)
5. Dyckman Street and Sherman Avenue (bus traffic) - On northeast corner (PM10)
6. 207th Street between 9th and 10th Avenues (truck traffic) - On the north side across from
Pathmark (PM10)
7. 214th Street and Broadway (bus and truck traffic) - On northwest corner (PM10)
CH-97-02
-------
CH-97-02
Figure 3. Site Locations for Phase 2
7
-------
Harlem Site
8. 125th Street and Amsterdam Avenue (high population density, shops, and bus and truck
traffic - same site is Phase 1 Site 3) - On northeast comer (PM10, PM10 quartz for
carbon analysis)*
* This site was the Harlem community site with the highest paniculate levels in Phase 1.
It was chosen to provide additional continuity and better assess general paniculate levels
between the two phases of the study.
At two community sites, Sites. 1 and 2, daily sampling for PM2.S was planned, but
logistical and operational problems resulted in only sporadic collection of PM2.5 data at the sites.
In addition, PM2.S data were collected at reference Site 9 on the reference sampler schedule of
approximately once every three days. Two additional portable samplers were operated daily in
Phase 2 (at the site carried over from Phase 1, Site 8, and at reference Site, 9) to collect samples
for analysis of elemental and organic carbon. Additional sampling at five sites was conducted
from September 21 through September 30, 1996. This additional sampling included daily
sampling for three community sites (Sites 2, 5, and 8), with sampling for elemental and organic
carbon at one site (Site 8). Every third day sampling continued at the two reference sites,
including sampling for elemental and organic carbon analysis at the midtown reference site
(Site 9).
During Phase 2, the PM10 reference samplers sampled on the following dates:
August 22, 25, 28, 31
September 3, 6,10, 12, 18, 24, 30
2.2 Sampling and Analysis
The following sections describe the paniculate samplers, field sampling procedures, and
analytical procedures for mass and carbon analysis.
2.2.1 Samplers
Saturation Samplers. The paniculate samplers used in the study were saturation samplers
supplied by EPA's SMR. The portable samplers are small, lightweight, and battery-operated;
they are ideal for monitoring in areas where it might be difficult and expensive to establish
permanent reference or equivalent samplers. The saturation sampler consists of a pump
controlled by a programmable timer which can be set to make up to six runs within a 24-hour
period. The portable miniVOL saturation samplers could be equipped with either a PM10 or
PM2.5 impactor inlet and is designed to sample at a flow rate of 5 liters per minute Opm) at
ambient conditions.1 Figure 4 is a diagram of a saturation sampler mounted on a utility pole.
The saturation sampler, however, is not an EPA reference or equivalent method for paniculate
matter.
CH-9742
-------
MMCLC
CH-97-02
Figure 4. Diagram of Mounted Saturation Sampler
9
-------
Reference Samplers. Reference samplers for PM10 were Sierra-Andersen Model 246B
dichotomous samplers, which have been officially designated by EPA as equivalent samplers to
the Federal Reference Method for PM10. These samplers sample at 16.7 1pm, have a PM10
inlet, and a 2.5 pm virtual impactor assembly. The sampler uses dual 37mm ringed-Teflon filters
for PM2.S and coarse (PM10-PM2.5). The sampler can thus provide measurements for PM10,
PM10-PM2.5, and PM2.5, but it is only designated as an equivalent method to the Federal
Reference Method for PM10 and not for PM2.5. Currently, there are no Federal Reference
Method or equivalent samplers for PMZ.5 because EPA has only proposed, and not promulgated,
a PM2.5 standard. These reference samplers were operated by New York State personnel. The
data from the reference samplers were obtained from the State by EPA Region D staff and used
to assess the relative accuracy of the PM10 and PM2.S saturation sampler data.
Sampler Placement Ideally, the sampler inlet for PM10 and PM2.5 sampling should be
at breathing height level (nominally between 1 and 2 meters). However, practical factors such
as prevention of vandalism, security, and safety precautions must also be considered. Given these
competing concerns, PM10 and PM2.5 sampler inlets are usually 2-7 meters above ground level.
As Figure 5 indicates, the saturation samplers were usually hung from lampposts and were
typically 3-4 meters above ground. Reference samplers were at ground level enclosed in
wrought-iron cages with inlets about 1-2 meters above the ground. Nonetheless, the saturation
samplers gave a good indication of the traffic contribution to PM10 and PM2.5. As discussed
in Section 3.4.3, at the midtown Manhattan reference site, good agreement was obtained for
PM10 values from the reference PM10 sampler located in the breathing zone and the saturation
PM10 sampler located on a lamppost
2.2.2 Sampling Procedures
The following sampling procedures, were used for this study:
• Two or more samplers per site per type of paniculate matter sampling (e.g., PM10,
PM2.5, or PM10 quartz) were generally hung from light or traffic signal posts.
• Samplers were set to run for 24-hours starting at midnight.
• Daily visits were made to sites that sampled daily.
• Each day a sampler of a particular type was reset for the next day while another sampler
of the same type was running.
• A flow calibration of the samplers was conducted at the start of the study and a check of
the flow calibrations was performed at the end.
x , i
• Flow points were set to achieve an actual ambient flow of 5 1pm every day prior to
sampling.
• Each sampler's flow rate was checked and adjusted, if necessary, prior to and after
sampling. '
/
CH-97-02 10
-------
9
3
6
(a)
(b)
Figure 5. Sampler Installations at (a) Site 1 and (b) Site 5 for Phase 1
-------
• Each sampler was checked for proper operation and damage including checks for the
following:
battery power
flow rate
elapsed time
leaks
unusual filter conditions (such as torn or discolored filters).
• Any abnormal conditions were noted on the field forms.
During both phases, 47mm Pallflex Teflon-coated glass fiber filters were used for
sampling PM10 and PM2.5 for gravimetric analysis. During Phase 2, 47mm Pallflex quartz
filters were used for sampling PM10 for analysis of elemental and organic carbon. Different
laboratories were used for the gravimetric and carbon analyses. Data for PM10 and PM2.5
gravimetric analyses were taken from the completed field data sheets and entered into a
computerized field data management package supplied by the analytical laboratory. At periodic
intervals the sampled filters, including field blanks, were sent along with copies of the field data
sheets and the field data on disk to the laboratory. The quartz filters for carbon analysis were
chosen after reviewing the gravimetric analysis results and sent to the analytical laboratory with
a summary data sheet in hardcopy and electronic formats.
2.23 Analysis Procedures
Mass. The following procedures were followed for determining mass by gravimetric
analysis:
• Filters for gravimetric analysis were pre- and post-weighed.
i \
• Prior to weighing, filters were allowed to equilibrate for at least 24 hours in a control box
where the relative humidity is below 50 percent (±5 percent) and the temperature is
constant to within ±3°C at 15°-30°C.
• Filters were weighed with a Cahn microbalance precise to ±1 ug.
• Just before weighing, filters were passed through the field of static eliminators for a few
seconds.
• Exposed filters were reweighed on the balance on which their tared weights were
obtained.
• A Class M 200ug weight certified by the State Meteorology Lab, and thus traceable to
National Institute for Standards and Technology (NIST), was used as a primary calibrating
standard.
• Tare of the balance was checked every 20 filters (readings should be 0.000 ±2pg).
CH-97-02 12
-------
• Calibration of the balance was checked at the start and end of each filter weighing session
using a 200pg Class M weight on the balance (calibration weight readings should be
±2pg of the 200pg weight).
• Three "standard" filters, arbitrarily chosen for the purpose, were weighed at the beginning
of each weighing session; if the filter's weight was not within 20pg of its established
value, a full-scale check-out of the balance was undertaken prior to regular filter
weighing.
• Every seventh filter was reweighed by the technician; if the weight was not within 20pg
of the original value, the problem was located and corrected, with all filters reweighed.
• Any blank filter weight outside the normal range of 55 to 65 mg resulted in immediate
investigation.
• Three "standard" filters from a batch were weighed with each batch of filters pre-weighed
ahd'post-weighed; if the average weight of these "standard" filters changed from the pre-
weighing to the post-weighing of filters due to changes in relative humidity, then a
correction factor was added to/subtracted from the post-weight value.2
Carbon Analysis. Quartz filters were analyzed for organic and elemental carbon using
a proprietary thermo-optical reflectance method. The following procedures were used:
• A section of the filter obtained by using a punch was inserted into an oven.
• The organic carbon was first volatilized at temperature steps between 300° and 550°C in
a 100% helium atmosphere and then combusted at temperature steps between 550° and
700°C in an atmosphere of 1-2% oxygen in helium.
• The carbon evolved at each temperature step was converted to methane by a methanator
and quantified in a flame ionization detector.
• The reflectance from the filter section was monitored using a laser from the beginning of
the process and throughout the process to correct for the pyrolysis of organic material.
Organic carbon (OC) is the carbon that evolves before the original reflectance is
reattained.
Elemental carbon (EC) is the carbon that evolves after the original reflectance is
reattained.2'3
• At the end of each run, methane from a calibration loop was injected for calibration and
diagnostic purposes. (The loop is calibrated by running known amounts of an organic
carbon standard such as potassium hydrogen phthalate.)
• The first three runs of each day consist of an instrument blank, a carbon standard, and an
EC/OC split reference standard of a previously characterized matrix deposited on a quartz
filter.
CH-97-02 13
-------
• Duplicate samples, were analyzed at a frequency of about one out of every 20 samples;
duplicate concentrations must not vary by more than 20% relative percent difference for
results greater than five times the detection limit. (Reanalysis was performed if the
duplicates did not meet this criterion.2)
Since the filters were not pre- and post-weighed, results were presented from the
: laboratory as micrograms per filter and were converted to ng/m3, based on the sample flow rate
using the algorithm given by the sampler manufacturer.
- 23 Operational Problems
i
Despite the procedures discussed in Section 2.4, some logistical, equipment, and
" operational problems occurred during the study that adversely affected data completeness and
data quality.- Although a pilot study was supposed to discover some of these problems, the
samplers used in the pilot study proved to be among the most reliable and did not hint at some
of the future problems. However, despite the problems encountered, the study was able to obtain
sufficient data to meet its objectives.
At the start of Phase 1, three problems surfaced: (1) some samplers did not keep their
original settings for 5 1pm from the field laboratory to the site locations, (2) some samplers were
not sampling for the elapsed time, and (3) it was difficult to get information on the status of
equipment and operations from the field operators. The first problem was quickly corrected after
it was discovered during the system audit.
The second problem of too short elapsed times was corrected in some cases by adjusting
the low flow cutoff threshold. However, most malfunctioning samplers could not be repaired.
They eventually were replaced with other, more reliable samplers as these other samplers became
available from the SMR. Considerable effort was expended in trying to diagnose and correct the
^problem. Defective batteries and operator error were eliminated as possible causes in a series
of tests. The problem was made more difficult to solve because the samplers would seem to
work in the field and SMR laboratories, but not at the site(s).
The third problem of coordination and communication difficulties was ameliorated by
providing the field operators with a long-distance calling card number which they could use to
report equipment problems and operations status 8 hours a day, 7 days a week.
The systems audit during Phase 1 revealed one other problem. The first sampler deployed
r at Site 2 was stolen on the first day of sampling. The field operators had relocated the
monitoring site from Broadway between 145th and 146th streets to a location one block east at
the northwest corner of Amsterdam Avenue and 145th Street. Site 2 was at this location from
July 13-25, 1996, but was returned to its original location on July 26.
CH-97-02 14
-------
2.4 Quality Assurance/Quality Control
A quality assurance project plan (QAPP) was prepared and approved prior to sampling.
The procedures contained in the QAPP were incorporated into standard operating procedures and
followed during the study.
A variety of quality assurance/quality control (QA/QC) procedures were followed for the
study.2 Sampling and analysis QC procedures and data quality indicators are discussed in
Sections 2.2 and 3.4. In addition, internal field QC checks included checks of battery power,
flow rate, sampled air volume, elapsed time, leaks, unusual filter conditions, and unusual site
conditions.
Field operators, who were undergraduate and graduate students at the City College of New
York, were trained for two days by a senior ambient air monitoring specialist in field operations,
equipment checks, field laboratory operations, recordkeeping, and site setup. A one-week pilot
phase with two sites was used as a test run to identify and resolve problems. Approximately
once a week, an experienced field technician would provide additional guidance to the students
who were under the supervision of a Professor of Meteorology at the school. Once during each
Phase, a senior QA manager conducted a systems audit of the operation and recommended and
implemented corrective action on the spot.
Data were reviewed on an ongoing basis and samples were invalidated if any of the
following conditions occurred:
• sampler flow rate >5.75 1pm or <4.25 1pm
• elapsed sampling time >30 hours or <18 hours
• sampled air volume >7.92 m3 or <6.48 m3
Data that did not meet other, more stringent criteria were labeled as suspicious. Thus, samples
were labeled as suspicious if any of the following conditions occurred:
• sampler flow rate >5.5 1pm or <4.5 1pm
• elapsed sampling time >24.5 hours or <23.S hours
PM10 S PM2.5
Other factors that would invalidate data included torn or damaged filters and holes in the filter.
Laboratory QC checks for gravimetric analysis included daily weighing of "standard"
filters, replicate weighings of every seventh filter, checks of balance tare, and calibration after
every fifth weighing. Laboratory checks for the carbon analysis included daily analysis of filter
blanks, carbon standard, and OC/EC split standard and a duplicate sample analysis at a frequency
of one per 20 samples or per batch, whichever is greater.
•
• Duplicate samplers were collocated at at least two sites during each phase to assess
overall precision, and one set of samplers was collocated at a reference sampler site to assess
relative accuracy. During the start of Phase 1, two sites had collocated PM10 sampling, Site 3
CH-97-02 15
-------
and reference Site 9, but collocated sampling was added at reference Site 10 on July 30, 1996,
to improve data completeness. All collocated sampling was chosen to correspond to the
enhanced reference sampling schedule of the State, which was approximately once every three
days. During the Stan of Phase 2, three sites had collocated PM10 sampling, Site 6 and reference
Sites 9 and 10, but collocated sampling was discontinued at Site 6 after September 20, 1996.
CH-97-02 16
-------
3.0 STUDY DATA
Data were obtained for PM10 and PM2.5 concentrations during both phases of the study.
In addition, data were obtained for elemental and organic carbon during Phase 2 of the study.
The following sections summarize the data obtained and their data quality indicators.
3.1 PM10 Data
Table 1 summarizes the results of the final data set for PM10 for Phase 1 and includes
daily summary statistics for percent complete and mean, minimum, and maximum values. Table
1 also provides similar summary statistics for each site for Phase 1. In addition, the table also
provides comments about weather conditions, such as rain and stagnation conditions, that could
affect PM10 levels. Table 2 provides the same information for the final data set for PM10 for
Phase 2.
The data in Table 1 show that during Phase 1, 24-hour PM10 levels ranged from a low
of 3 pg/m3 (a suspicious value) at Site 6 (the control site) to a high of 122 pg/m3 at Site 9 (a
reference site). This high value was due to construction at the site and confirmed by several
samplers. In general, if the exceptional value for Site 9 is omitted, Site 10, the reference site in
downtown Manhattan had the highest average PM10 concentrations, 54 pg/m3, and Site 9, the
reference site in midtown Manhattan, had the next highest average PM10 concentration, 52
pg/m3. Site 3 was the Harlem community site with the highest average PM10 concentrations,
but its average concentration for the period, 40 pg/m3, was about three-fourths that for the
reference site, Site 10. Site 3 was located at the northeast corner of 125th Street and Amsterdam
Avenue, a site with high population density, shops, and bus and truck traffic. Site 6, the control
site, had the lowest average PM10 concentration, 28 pg/m3. Figure 6 shows mean PM10
'concentrations for Phase 1 for all days, weekdays, and weekends.
Table 2 shows that during Phase 2, 24-hour PM10 levels ranged from a low of 12 pg/m3
(a suspicious value) at Site 4 (the control site) to a high of 79 pg/m3 at Site 9, the midtown
reference site. In general, Site 10, the downtown reference site, had the highest average PM10
concentrations, 52 pg/m3, and Site 9, the reference site in midtown Manhattan, had the next
highest average PM10 concentration, 46 pg/m3. Site 7 was the Washington Heights community
site with the highest average PM10 concentration, 36 pg/m3, but its average concentration for the
period was about two-thirds that for the reference site, Site 10. Site 7 was located at the
northeast corner of 214th Street and Broadway, a site with heavy bus and truck traffic. Site 5
had the lowest average PM10 concentration, 22 pg/m3, while Site 4, the control site, had the next
lowest average PM10 concentration, 29 pg/m3. However, due to logistical and operational
problems, the sampling periods for Sites 4 and 5 did not overlap. Figure 7 shows mean PM10
concentrations for Phase 2 for all days, weekdays, and weekends.
Examination of Tables 1 and 2 shows that PM10 concentrations may vary from site to
site and from day to day. During the stagnation period of June 26 through June 29, PM10
concentrations were above average at all sites. Thus, average concentrations may be misleading
when different sites may have sampled for different days. Site rankings were used in an initial
attempt to compensate for the difficulties of different sampling days. Tables 3 and 4 summarize
CH-97-02 17
-------
Tablet PMlODau(taug/m3) tor PhaMl-Harlwn Community
OavNc
1
2
3
10
11
12
13
14
16
16
17
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Oar
Wed
Thu
Fn
Sat
Sun
Mon
Tua
Wad
Thu
Fn
Sa)
Sun
Mon
Tua
Wad
Thu
Fn
Sat
Sun
Mon
Tue
Wad
Thu
Fn
Sal
Sun
Mon
Tua
Wad
Thu
Fn
Sat
Sun
Mon
Tue
Wad
Thu
Fn
Sat
Sun
Man
Date
07/10
07/11
07/12
07/13
OT/14
07/16
07/16
07/17
07/18
07/19
07/20
07/21
'07/22
07/23
07/24
07/25
07/26
07/27
07/28
07/29
07/30
07/31
08/01
06/02
06/03
06/04
06/05
06/06
08/07
08/06
06/09
08/10
08/11
08/12
06/13
06/14
08/15
06/16
06/17
06/16
08/19
I
I
1
18
33
17
33
41
39
39
48
22
12
6B
33
60
34
11
22
37
26
27
61
69
78
77
79
M
fgT
»
15
14
26
40
34
23
27
26
i
•
2
29
39
13
26
38
61
63
40
48
33
25
24
20
44
38
27
47
29
|
«
3
26
30
40
16
43
46
49
19
10
35
40
63
68
36
„..
31
60
67
66
89
47
48
23
19
42
29
44
30
30
49
29
32
s
I
4
13
22
16
43
29
33-
22
31
42
38
26
15
17
36
20
16
26
'63
64
62
70
36
19
14
26
„
36
31
25
26
24
23
,
«
I
6
24
17
44
42
47
25
11
33
47
61
34
17
20
35
29
18
35
63
62
76
78
78
41
60
18
18
34
46
26
I
|
6
16
34
22
29
38
21
12
cm
41
25
16
17
30
18
16
| 3 ]
60
64
47
64
69
34
34
21
13
30
30
22
20
25
22
1
}
7
25
31
20
36
35
36
44
24
11
26
43
65
16
SO
35
22
22
39
64
64
65
86
82
41
44
16
27
23
36
36
23
47
22
27
I (
I \ I
6 9 10
23
45
49
21 14
42
71
62 63
38 66
7
1 65~l44
69
44 164 169
68
93 94
43
47
122 31
12
12
29 39 61
43
39 62 65
27
57
32
32 44 72
Nad
6
4
3
6
4
6
6
6
6
6
6
4
4
7
7
4
7
6
6
6
7
7
6
6
6
8
3
7
8
7
8
'7
8
6
8
7
10
7
7
6
10
Naxp
10
6
6
8
10
8
8
10
8
6
10
6
6
10
8
8
10
6
6
8
10
8
10
6
8
10
8
8
10
8
8
10
6
8
10
8
10
6
8
6
10
* Compl
600
600
376
626
400
625
626
600
760
625
600
600
600
700
875
600
700
625
625
625
700
876
600
750
625
600
375
875
800
675
1000
700
1000
750
800
875
1000
675
675
750
1000
Mean
189
254
347
172
370
369
341
406
442
228
155
11 1
363
313
496
630
371
169
190
343
299
193
342
645
61 '0
687
74.6
803
832
403
432
364
164
243
296
390
379
253
41 9
264
333
Mm
128
22
307
159
333
221
285
325
382
212
106
96
26
22
413
375
247
11 2
166
299
163
67
3
601
641
436
701
635
694
342
288
185
116
122
192
298
297
22
204
21 7
219
Max
253
30
402
197
428
444
457
49
487
246
206
119
677
421
709
679
53
179
221
366
663
313
651
602
692
69
783
66
937
465
60S
122
248
424
605
44
55
301
S6B
31 6
716
Wejathor Comments
Frontal passage
Rain
Frontal passage
•
Rain
- TT
Rain
Stan ol stagnation penod
Stagnation
Stagnation
Stagnation
Stagnation
Stagnation
End of stagnation penod
Rain
Ram
'
MDays Nad 35 18 31 31 29 31 34 18 12 10 I (Suspicious data
N»xp 41 41 41 41 41 41 41 41 14 14
%Compie54 439 766 766 707 766 829 439 657 71 4
Mean 367380406303390283364379680638
Mm 11 13 96 13 11 22 11 67 21 14
Max 7683697078698866 122 94
Rain noted when >« o 25 in
Nact 26 14 22 22 21 21 24 14 9 7
Nexp 2929292929292929 10 10
UCompI 89 7 48 3 769 769724 724828483 900700
39340646031541.7294383384643606
Mm 14 20 25 13 16 22 22 67 38 44
Max 79 63 69 70 78 69 66 66 93 94
Nact 94998 10 10 4 3 »
Nnp 12 12 12 12 12 12 12 12 4 4
%Comp)750 333 750 760 667 833 633 333 760 76.0
262294294 274 321 261 31 936 9690380
11 13 96 14 11 12 11 12 21 14
69 47 67 62 63 64 65 67 122 69
Mm
Max
Parcert CompMa (% Compl). irxr(Adual Numbar (NactyExpeaed Number [Nexp])
CH-97-02
18
-------
Table 2 PM10 Data (taugma) tor Phaw 2 •WaihmgtonHa«htt Community
DavNo Day
1
2
3
4
6
6
7
8
9
to
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Thu
Fri
Sat
Sun
Mon
TIM
Wed
Thu
Fri
Sat
Sun
Mon
TIM
Wed
Thu
Fn
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sat
Sun
Mon
Date
0642
0843
0844
0645
0646
0847
0848
0849
0800
0801
09/01
09/02
09/03
09/04
09/05
09/06
09/07
09/08
09109
09/10
09/11
09/12
09/13
09/14
09/15
09/16
09/17
09/18
09/19
0940
0941
0942
0943
09/24
09/25
0946
0947
0948
0949
09/30
«
«
•
1
37
31
42
43
43
30
28
27
43
64
64
36
S?
35
39
47
24
20
17
17
25
31
22
26
24
!
I
2
34
30
68
36
37
48
21
24
38
25
27
37
33
29
35
17
23
27
23
31
1««h*9lNfchota
162nU EdBMomb*
Dydtmena SMniwn
346
48 66
60
31
27
37 40
48
36
35 32
27 26
»
28 22
40
64
67
33 32
32
31 31
37
44
24
23
19
14 16
22 22
60
14 12
22 21
15
30
18
25
16
16
22
27
21
1
6
61
67
33
26
41
41
37
30
26
24
27
22
36
67
62
34
32
41
34
22
20
16
16
26
27
13
20
19
|
K
7
63
66
35
27
43
47
38
34
29
26
39
64
33
43
22
47
18
18
23
23
1 I I
8 9 10
61 76 71
68
34
29 34 37
44
44 61 47
40
29
23 34 38
28
25
42 60 68
67
49 47 76
40
39
44
60 79
33
32 36 61
31
21
27
33
20[ 22)
25
29
29
33
24
33 42 49
25
22
19 36 60
37
26
27 37 38
Nad
7
4
6
8
6
4
6
6
4
7
6
6
7
4
6
7
c
6
6
6
5
6
6
7
6
6
0
7
6
6
2
3
3
4
3
3
4
3
3
6
Nexo % Comol
10
8
6
10
8
8
10
8
6
10
8
8
10
8
8
10
8
8
8
10
6
10
8
8
10
8
8
10
8
10
3
3
3
6
3
3
6
3
3
5
700
600
626
800
625
600
600
760
600
700
626
750
700
600
626
700
S26
750
626
600
62S
600
625
875
600
625
00
700
750
600
667
1000
1000
800
1000
1000
800
1000
1000
1000
Mean
694
627
340
306
408
443
438
356
318
290
276
259
444
630
646
439
352
341
407
601
250
325
206
183
241
357
163
229
226
326
316
234
372
256
186
320
284
254
306
Mm
485
669
313
264
371
406
359
298
263
232
256
221
384
667
62
316
322
306
373
341
22
197
166
136
216
266
124
201
147
293
297
179
251
163
163
193
216
228
206
Max Weather Comments
769
664
369
371
438
479
61 1
431
432
38
284
304
684
676
669
762
396
388 Ram
435
785
331
61 3
308
243
265
497
Rain
25 Ram
256
293
365
326 Ram
287
466
351
216
495
37
273
379
AD Days Nad 25 20 27 12 8 28 20 37 12 10 I ISuapooui data
Nexp 3040303040303040 14 14
%Ccmpl63 3 600 900 400 200 93 3 667 926 65 7 714
Maan 337321338265219316358343460624
Mm 17 17 14 12 16 13 18 19 22 37
Max 6468646630 67 666679 75
Ram notad whan >• 0 26 m
Waattay* Nad IB 13 21 8 6 21 16 25 10 6
Nexp 2228222228222228 11 11
HCompiei 6 464 955 364 179 966 682 893 909 727
Maan 354329367307193332396362466661
Mm 17 17 14 12 16 13 16 19 22 38
Max 64 66 64 66 26 67 66 68 79 75
Nad 8776366 12 22
Nexp 8 12 12 12 12 12 12 12 3 3
UCompHOO 663663 41 7 25.0 667 600 100 667667
Maan 300306269263262276269304339376
Mm 17 23 14 16 22 16 18 21 34 37
Max 37 37 32 31 30 34 35 40 34 38
Percent Complete {% CompO « 100*(Actual Number [NactJ/Expeaed Number [Nexp])
CH-97-02
19
-------
9
I
Manhattan Community Based Particulate Study
Mean PM10 Concentrations by Site for Phase 1
Influence of one point
08/10/96-122 ug/m3
Alt Days
Weekdays
Weekends
She 1:133rd & Broadway, Sfte2:145th & Broadway, Site 3:125th & Amsterdam, Site 4: Lennox & 148th. Site 5:135th & Lennox,
Figure 6. Mean PM10 Concentrations by Site for Phase 1
-------
Manhattan Community Based Particulate Study
Mean PM10 Concentrations by Site for Phase 2
9 10
• All Days
M Weekdays
• Weekends
Sites ~ 1:181st & St. Nicholas; 2:179th & Broadway; 3:168th & St. Nicholas; 4:162nd & Edgecombe (Control); 5: Oyckman & She
6:207th & 9th Ave.; 7:214th & Broadway; 8:125th & Amsterdam (Phase 1 Site 3): 9: Madison Ave. (Ret.); 10: Canal St. (Ret.)
rmal
Figure 7. Mean PM10 Concentrations by Site for Phase 2
-------
Table 3. Phase 1 Harlem Community Site Rankings
Day No
1
2
3
4
5
6
7
8
9
. 10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
36
39
40
41
Day
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
N
Mean
Date
07/10
07/11
07/12
07/13
07/14
07/15
07/16
07/17
07/18
07/19
07/20
07/21
07/22
07/23
07/24
07/25
07/26
07/27
07/28
07/29
07/30
07/31
08/01
08/02
08/03
08/04
08/05
08/06
08/07
08/08
08/09
08/10
08/11
08/12
08/13
08/14,
08/15
08/16
08/17
08/18
08/19
Relative Rank
i
1
3
2
3
4
3
2
4
2
4
1
1
5
3
5
5
1
1
4
2
5
1
1
6
6
6
8
6
5
5
5
4
6
5
3
8
35
3.9
8
|
2
3
5
4
1
4
3
4
4
5
2
2
1
4
7
1
4
2
5
5
18
36
6
i
3
1
1
1
4
1
1
1
2
4
2
2
2
1
3
1
1
6
3
3
1
3
4
2
1
3
1
10
1
2
2
3
31
2.4
3
I
4
5
4
4
2
4
6
3
6
6
4
6
4
4
4
6
6
5
4
5
5
3
6
7
6
3
5
8
4
6
5
9
31
5
9
i
5
3
2
1
3
3
1
3
4
4
3
4
3
2
2
3
4
4
3
4
2
5
7
3
1
3
8
7
3
7
29
3.5
5
1
w
6
4
3
5
4
6
5
2
7
7
7
1
5
5
7
5
6
6
4
7
7
8
7
7
5
7
7
9
7
7
4
to
31
5.8
10
I
7
2
3
1
2
4
5
4
2
5
3
5
2
2
3
3
5
3
3
3
2
2
1
5
3
5
4
2
6
6
5
6
4
6
6
34
36
7
i
8
2
1
1
7
2
8
2
2
4
8
6
3
3
3
3
1
1
4
18
3.4
4
I
9
2
1
1
2
2
1
3
2
1
2
2
2
12
1.8
2
I
10
3
1
1
2
1
1
3
1
1
1
10
1.5
1
N
5
4
3
5
4
5
5
6
6
5
5
4
4
7
7
4
7
5
5
5
7
7
6
6
5
8
3
7
8
7
8
7
8
6
8
7
10
7
7
6
10
CH-97-Q2 22
-------
Table 4. Phas? 2 Washington Heights Community Site Rankings
Day No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
8/22-9/30
Day
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Prl
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
,Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
N ,
Mean
Date
08/22
08/23
08/24
08/25
08/26
08/27
08/28
08/29
08/30
08/31
09/01
09/02
09/03
09/04
09/05
09/06
09/07
09/08
09/09
09/10
09/11
09/12
09/13
09/14
09/15
09/16
09/17
09/18
09/19
09/20
09/21
09/22
09/23
09/24
09/25
09/26
09/27
09/28
09/29
09/30
Relative Rank
1st &St Nicholas
co
1
1
4
3
1
-1
3
1
3
3
2
3
5
2
3
4
4
3
8
5
5
3
4
2
1
3
,
25
3.1
5
«c
v»
2
2
1
1
4
2
3
6
1
2
1
2
1
1
1
1
2
3
2
3
3
20
2.1
3
Bin A SL Nicnolfts 1
*
3
7
3
6
5
1
6
3
4
1
2
5
3
1
6
5
5
5
5
2
5
2
7
6
1
5
4
5
27
4.1
8
|
•8
1
«£
4
4
5
4
5
5
5
7
6
6
5
7
5
12
5.3
10
*
£)
5
3
3
4
3
3
3
1
5
B
3.1
6
Tin & 9th Ave. 1
8
6
6
4
4
8
3
4
5
6
4
6
4
6
7
4
5
3
4
3
6
5
7
4
3
2
5
6
6
4
28
4.8
9
•e
f
N
7
5
2
2
7
2
2
4
4
3
5
6
3
4
2
4
2
3
3
4
3
20
3.5
7
5th & Amsterdam I
CM
*•
8
3
1
3
5
1
3
2
2
7
3
4
4
2
2
1
1
1
2
1
4
1
2
1
3
4
2
1
2
1
2
3
2
1
4
1
2
4
37
2.4
4
idison Ave.
2
9
1
3
1
" "
2
2
3
1
3
3
2
2
2
12
2.1
2
55
0
10
2
1
2
1
1
1
1
1
1
1
10
1.2
1
•
N
7
4
5
8
5
4
6
6
4
7
5
6
7
4
5
7
5
6
5
6
5
8
5
7
6
5
0
7
6
5
2
3
3
4
3
3
4
3
3
5
CH-97-02
23
-------
the results of the site rankings for Phases 1 and 2, respectively. On any given sampling day, the
site with the highest 24-hour PM10 concentration was given a rank of 1, the next highest a rank
of 2, and so on. The average rank for each site during the entire study period was used to
•determine an overall relative rank. In each phase, Site 10 was ranked 1st and Si'e 9 was ranked
2nd. During Phase 1, Site 3, ranked 3rd, was the Harlem community site with the highest
relative ranking, and Site 6, the Harlem community control site, was the community site ranked
last. During Phase 2, Site 2, ranked 3rd, was the Washington Heights community site with the
highest relative ranking, and Site 4, the Washington Heights community control site was ranked
last. For Phase 2, Site 8 (which was Site 3, the highest ranked community site with a rank of
3rd, during Phase 1) was ranked 4th.
In general, average PM10 concentrations in Phase 2 were down slightly from those in
Phase 1 - about 4 percent at Site 10, 12 percent at Site 9 (excluding the exceptional day), and
15 percent at Site 8 (Site 3 in Phase 1). Appendix A contains a complete set of plots of daily
PM10 concentrations for each site during Phase 1. Appendix B contains the sample plots for
Phase 2.
3.2 PM2.5Data
During Phase 1, PM2.5 samplers were located at Sites 1, 5, and 9. During Phase,2,
PM2.5 samplers were located at Sites 1,2, and 9. Tables 5 and 6 summarize the data for PM2.S
during Phases 1 and 2, respectively. The tables show both PM10 and PM2.S concentrations in
.ug/m3 and the ratio of PM2.5/PM10 at each site for each day. Daily summary statistics for
percent complete and mean, minimum, and maximum ratios are also given. In addition, similar
statistics are provided for PM10, PM2.5, PM2.5/PM10 for each site for all days, weekdays, and
weekends for each phase. Figure 8 shows daily PM10 and PM2.5 concentrations for Site 1
during Phase 1. Appendix A also shows daily PM2.5 concentrations along with PM10
concentrations for all sites having PM2.5 sampling during each phase. In addition, PM2.5 was
^inadvertently acquired at Site 3 for three days (July, 14, 17, and 24) during the start of Phase 1
when a PM2.5 sampler inlet was erroneously used on what was supposed to be a collocated
PM10 sampler.
Table 5 shows that Phase 1 24-hour PM2.5 levels ranged from a low of 2 ug/m3 at Site
5 to a high of 79 ug/m3 at Site 9 (the midtown Manhattan reference site). During Phase 1, mean
ratios of PM2.5/PM10 ranged from 0.75 at Site 5 to 0.83 at Site 9. Table 6 shows that Phase
2 24-hour PM2.5 levels ranged from a low of 12 ug/m3 at Site 2 to a high of 61 ug/m3 at Site
9. Logistical and operational problems reduced the amount of PM2.5 data that were acquired
during Phase 2.
During Phase 1, mean ratios of PM2.5/PM10 ranged from 0.75 at Site 5 to 0.83 at Site 9.
During Phase 2, mean ratios ranged from 0.69 at Site 1 to 0.77 at Site 2. The mean PM2.5
concentration at Site 9 decreased from 47 ug/m3 during Phase 1 to 35 ug/m3 during Phase 2
while the mean ratio of PM2.5/PM10 varied from 0.83 during Phase 1 to 0.75 during Phase 2.
Figures 9 and 10 show daily PM2.5 concentrations along with PM10 concentrations at Site 9 for
Phases 1 and 2, respectively.
CH-97-02 24
-------
TabtoS. PM2 5 Ctata (In u»^) tor PhaMl-Haftom Community
Day No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
All Days
SHe1
J 193rd I Broadway
Day Date (10) (25) (25/10)
Wed 07/10 18 17 0.983
Thu 07/11 18
Fri 07/12 33
Sat 07/13 17 11 0.671
Sun 07/14 33
Mon 07/15 41 30 0.723
To* 07/16 39
Wed 07/17 39 32 0.818
Thu 07/18 48 38 0.786
Fri 07/19 22 18 0.817
Sat 07/20 11
Sun 07/21 12 9 0.765
Mon 07/22 58 25 0.435
To* 07/23 33 28 0.848
Wed 07/24 50 44 0.878
Thu 07/25 44
Fri 07/26 34 29 0867
Sat 07/27 111 111 11.009
Sun 07/28 22
Mon 07/29 37 31 0844
To* 07/30 26 21 0801
Wed 07/31 27 23 0.857
Thu 08/01 33
Fri 08/02 51
Sat 08/03 69 61 0.874
Sun 08/04 56
Mon 08/05 78
To* 08/06 77 68 0.88
Wed 08/07 79 66 0.838
Thu, 08/08 38 31 083
Fn 08/09 1 29 I 30 11035
Sat 08/10 20 16 0.821
Sun 08/11 15 12 0799
Mon 08/12 14
Toe 08/13 28 15 0521
Wed 08/14 40
Thu 08/15 34 27 0798
Fri 08/16 23 18 0779
Sat 08/17 31
Sun 08/18 27 24 0.892
Mon 08/19 26 20 0781
Nad 35 33 27
Nexp 41 41 41
%Compl854605 659
Mean 36 29 0.813
Mm 11 91 0.435
Max 79 68 1.035
Sit* 5
136th I Lennox
(10) (25) (2.5/10)
24
17
44
42
47
25
11
33
47
51
34
17
20
35
29
18
35
63
62
75
78
78
41
60
18
19
34
48
26
29
41
707
39
11
78
13
24
11
18
28
30
34
9
11
15
39
42
26
13
17
27
15
12
26
41
49
55
64
68
59
31
48
14
8
14
42
24
2
38
16
19
36
41
87.8
27.8
2.3
681
0.649
0.631
0.727
0.956
0519
0814
0759
0764
0.838
0.756
0533
0.678
0744
0.777
0875
0851
0872
0.764
0751
0.805
0472
0.714
0.714
0.791
0720
25
41
61.0
0.761
0472
0.956
Stte9
MadaonAve
(10) (25X25/10)
45
21
71
52
38
HE
| 64
93
122
39
52
44
12
14
857
58
21
122
16
67
44
26
IM:
| 68
79
41
31
45
37
11
14
786
47
16
79
0.782
0.805
0.849
0702
] 1.232
1 1063
0.846
0.339
0.782
0.861
0828
11
14
78.6
0.826
0.339
1232
PM25/PM10
NadNexpSCompI
1
0
0
2
1
e
0
2
2
1
1
2
1
1
4
1
3
2
1
2
3
2
2
0
2
2
, 1
2
3
2
2
2
2
0
3
0
3
1
1
1
3
I |St
•j^BVwJ
3
2
2
2
3
2
2
3
2
2
3
2
2
3
2
2
3
2
2
2
3
2
3
2
2
3
2
2
3
2
2
3
2
2
3
2
3
2
2
2
3
tmrtlI
npw
33.3
00
00
100.0
333
100.0
00
66.7
1000
500
33.3
1000
500
33.3
2000
50.0
1000
1000
50.0
1000
1000
1000
667
0.0
1000
667
500
1000
100.0
1000
1000
66.7
1000
00
1000
00
1000
50.0
50.0
500
1000
Diou* 29 29 29 29 29 29 10 10 10
%Compl89.779.3 69.0 72.482.8 586 90.080.0 80.0
Mean 39 31 0.806 42 309 0.744 54 46 0863
Mm 14 15 0.435 18 2.3 0533 38 26 0.702
Max 79 68 1.035 78 681 0.872 93 79 1232
WMkand* Nad 9 10 7 8 12 8 333
N*xp 12 12 12 12 12 12 444
%CompJ75.083.3 58.3 66.7100.0 66.7 75.075.0 75.0
Moan 25 24 0.833 32 21.6 0.765 69 42 0.728
Min 11 91 0.671 11 8.3 0.472 21 16 0.339
Max 69 61 1.009 63 546 0.956 122 68 1.063
P*ro»nt Compute (% CompO - 100'(Adual Number [NadyExp*ct»d Number [NexpJ)
CH-97-02
25
-------
TabtoS PM25DaU (in up/m3) for Phaaa 2-Waahinp^Haigrits Community
Day No. Day Data
1 Thu 08/22
2 Fri 08/23
3 Sat 08/24
4 Sun 08/25
5 Mon 08726
€ Tu« 0877
7 Wad 0878
8 Thu 0870
0 Fri 0800
10 Sat 08/31
11 Sun 09/01
12 Mon 0*02
13 Tua 08/03
14 Wad 08/04
15 Thu 08/05
16 Fit OS/06
17 Sat 09/07
18 Sun 09/08
19 Mon 09/09
20 TIM 09/10
21 Wad 09/11
22 Thu 09/12
23 Fri 09/13
24 Sat 09/14
25 Sun 09/15
26 Mon 09/16
27 Tu« 09/17
28 Wad 09/18
29 Thu 09/19
30 Fn 09/20
31 Sat 09/21
32 Sun 09*72
33 Mon 09/23
94 TIM OV24
35 Wad 09/25
36 Thu 09/26
37 Fri 09/27
38 Sat . 0078
39 Sun 09*9
40 Mon 09/30
All Days Nad
Naxp
Sttal
181(1 »St Mcnolas
(10) (2 X) (2 5/10)
38
37
31 22
42 23
43 33
43
30
28
27
43
64
54
35
37
35
39
47
24
20
17
17
25
31
1 22l 171
26
24
25 5
30 30
%Compl83.3 167
Maan
Mm
Max
Waakdaya Nad
Naxp
34 27
17 17
64 38
17 4
22 22
%Compl77.3 18.2
Maan
Mm
Max
Waakanda Nad
Naxp
85 28
17 17
64 38
8 1
8 8
KCompMOO 125
Maan
Mm
Max
80 22
17 22
37 22
0.706
055
0.761
| 0756
4
30
13.3
0.693
055
0.761
3
22
13.6
0.689
055
0.761
1
8
125
0.706
0.706
0.706
Sits 2
1 79th t Broadway
(10) (25) a 5/10)
34
30
68
36
37
48
21
24
38
25
27
37 25
33
29
33
35
17 19
23 12
27
23 17
31
25 5
40 40
625 125
32 21
17 12
68 33
13 3
28 28
46.4 10.7
33 22
17 12
68 33
7 2
12 12
S8.3 16.7
31 21
23 17
37 25
'
0679
111
0534
0.75
4
40
10.0
0768
0534
1.11
2
28
7.1
0.822
0534
1.11
2
12
16.7
0.715
0.679
075
Sit* 9
Madison Ava
no) (25) (2.5/10)
76 61
34, 27
61 45
34 23
50 40
47 38
79 60
36 29
I 22! 181
42 16
36 27
37 32
12 12
14 14
85.7857
46 35
22 16
79 61
10 10
11 11
90.9909
48 37
22 16
79 61
2 2
3 3
66.7 66.7
34 25
34 23
34 27
0.805
0.807
0741
0.672
0.806
0.797
0.761
0.806
| 0.811
037
0758
0872
12
14
857
0.75
037
0.872
10
11
90.9
0753
0.37
0.872
2
3
66.7
0.739
0.672
0807
PM25/PM10
N*tNaxp%Compl
1
0
0
2
0
1
1
1
0
1
0
0
1
0
0
1
0
0
0
1
0
1
0
0
0
0
0
2
0
0
1.
0
0
1
0
1
2
0
1
1
3
2
2
3
2
2
3
2
2
3
2
2
3
2
2
3
2
2
2
3
2
3
2
2
3
2
2
3
2
3
1
1
1
3
1
1
3
1
1
3
333
00
00
66.7
00
500
333
500
0.0
33.3
0.0
00
333
0.0
0.0
33.3
00
0.0
0.0
333
00
33.3
00
00
00
00
0.0
667
00
0.0
1000
0.0
0.0
333
00
1000
66.7
0.0
1000
333
PM25/PM10
Maan Mm Max
081
0.76
055
0.74
0.76
0.67
081
08
0.76
0.8!
0.78
0.68
037
1.11'
065
075
087
081 081
0.71 0 81
055 055
074 074
076 0.76
0.67 067
.
081 0.81
0.8 0.8
076 0.76
081 .081.
0.76 0.81
068 068
037 037
111 111
053 076
075 075
087 087
I ISutpioom data
Parcant Compiata (% Compi) • 100*(Aotual Number [NactyExpactad Numbar (Naxp])
CH-97-02
26
-------
V ft.
Manhattan Community Based Particulate Study
Daily PM10 and PM2.5 Concentrations for Phase 1 Site 1
160
120
I 80
0)
o
8 40
10
20 30
Sampling Day
PM10
PM2.5
40
50
Site 1:133rd & Broadway |
Figure 8. Daily PM10 and PM2.5 Concentrations at Phase 1 Site 1
-------
9
£
fe
oo
Concentration (ug/m3)
4^ oo ro oo
o o o o o
*^
Manhattan Community Based Particulate Study
Daily PM10 and PM2.5 Concentrations for Phase 1 Site 9
1 •
I •
^X KJ
H O /s • v |
i i i i
• PM10
• PM2.5
-
) 10 20 30 40 50
Sampling Day
Site 9: Madison Ave. (Ref .) |
Figure 9. Daily PM10 and PM2.5 Concentrations at Phase 1 Site 9
-------
K>
Concentration (ug/m3)
l\> 4fc O> 00 O
o o o o o o
*•» .
Manhattan Community Based Particulate Study
Daily PM10 and PM2.5 Concentrations for Phase 2 Site 9
•
• •
* • 0
* "•
" • • * * • • *
" ^" ^
• ^ O
^ •
" ®
i i i - i
• PM10
<*> PM2.5
) 10 20 30 40 50
Sampling Day
| Site 9: Madison Ave. (Ref.)|
Figure 10. Daily PM10 and PM2.5 Concentrations at Phase 2 Site 9
-------
3.3 Carbon Analysis Data
During Phase 2, PM10 samplers equipped with quartz filters were located at Sites 8
(Phase 1 Site 3) and 9 to obtain measurements of elemental and organic carbon [which together
constitute total carbon (TC)] using a thermo-optical analysis method.2-3 Daily samples were
collected at Site 8 while collocated samples were taken at Site 9 on the reference sampler
schedule of about once every three days. Table 7 summarizes the PM10 carbon analysis data and
gives the daily measurements of PM10, EC, OC, and TC in pg/m3 and the ratios EC/OC, EC/TC
TC/PM10, and EC/PM10. In addition, it provides summary statistics for each site for all days,
weekdays, weekends, and all paired values. Table 7 shows that the highest concentration of
carbon occurred at Site 8 - about 8.1 pg/m3 for OC, 19.3 pg/m3 for EC, 25.5 pg/m3 for TC. The
ratio of EC/PM10 varied from a low of about 0.05 to a high of about 0.66 at Site 8. In general,
EC was about 70 percent of TC, TC was about 50 percent of PM10, and EC was about 30
percent of PM10.
3.4 Data Quality Indicators
Two major indicators of data quality were used for study data, precision and relative
accuracy. Precision is the degree of mutual agreement among individual measurements under
prescribed conditions. Two aspects of precision were examined: overall precision obtained from
collocated PM10 samplers and analytical precision obtained from replicate analyses of samples.
Analytical precision was found to be better than overall precision.
3.4.1 Overall Precision
During Phase 1, collocated data were obtained at Sites 3, 9 and 10. Table 8 summarizes
the collocated data for Phase 1. Table 8 shows that for individual sampling days, collocated
precision error varied from a low of about 0.6 percent to a high of about 28 percent. The
» average precision error at a given site varied from a low of 1.9 percent with a standard deviation
* of 9.8 at Site 3 to a high of about 9.9 percent with a standard deviation of about 14.8 at Site 10.
For all of Phase 1 the average precision error at all collocated sites was about 0.3 percent with
a standard deviation of about 12.1. If statistics are calculated as defined in the Quality Assurance
Project Plan2 and 40 CFR Part 58, Appendix A,4 then the 95 percent upper probability limit is
23.9 percent and the 95 percent lower probability limit is -23.4 percent.
During Phase 2, collocated data were obtained at Sites 6, 9, and 10. Table 9 summarizes
the collocated data for Phase 2. Table 9 shows that for individual sampling days collocated
; precision error varied from a low of about 0.4 percent to a high of about 15.1 percent. The
average precision error at a given site varied from a low of 1.7 percent with a standard deviation
of 6.2 at Site 10 to a high of about 5.8 percent with a standard deviation of about 8.8 at Site 6.
For all of Phase 2, the average precision error at all collocated sites was about -2.8 percent with
a standard deviation of about 5.3. The 95 percent upper probability limit is 7.6 percent and the
95 percent lower probability limit is -13.3 percent
CH-97-02 30
-------
Tab* 7 PM10C*rbonAmlyMCMa(*) 119*13) tor Phaa* 2
Day No
1
2
3
4
6
6
7
a
8
10
11
12
13
14
15
16
17
16
19
20
21
22
23
24
25
26
27
26
29
30
31
32
33
34
35
36
37
38
39
40
aVTUVQ/Vt t
W9^Wxp
%Compl
Mun
Mn
Max
sNacl
PM10OC
61
68
94
29
44
44
40
29 237
23 3*0
28 4*0
25 424
42 042
67 636
48 370
40 148
38 218
44 4*0
60 1*7
93 622
22 367
31 2*6
21 1*5
27 431
93 2*0
20 221
26 7*5
28 476
29 611
33 7*2
24 236
33
25
22 1*5
19 3*6
22
26 3*6
27 445
30 27
32 32
83* 644
31 3*
18 04
68 61
20 18
s«*e
EC/
EC TC OC
638 676270
611 10* 1*7
630 132 1*9
635 10* 1*0
1*3 235 4*4
122 197 1*1
13* 17*376
474 621 320
6*9 7*8 2*9
10* 166 225
4*6 6*2 2*1
193 26*310
106 142 2*6
166 196 6*6
735 920 3*7
806 134210
601 10*2*6
744 8*5 337
162 233 230
130 177273
122 203 1*0
11* 184 1*5
716 8*2301
5*8 7*3 301
8*0 134 247
611 818 1*9
124 16*277
27 27 27
32 32 32
844 844 644
84 13 27
1* 24 1*
19 25 6*
16 18 16
Sla9Sampl*rA
EC/ TC/
EC/
TC PM10PM10
073 030
0*1 043
0*3 047
060 042
0*2 0*6
0*2 035
078 036
076 016
072 020
0*8 036
074 013
076 077
074 0*3
0*5 0*4
0*0 045
0*8 0*0
074 032
077 048
070 0*4
073 061
0*0 0*8
0*1 060
075 040
076 036
071 0*8
067 035
074 061
97 97
f.f «*
32 32
644 844
07 0*
06 01
06 0*
18 16
022
026
030
025
0*6
022
026
012
016
025
010
0*6
047
0*4
036
034
024
036
066
044
042
036
030
027
048
023
045
97
C(
32
644
03
00
07
18
ED
PM10OC EC TC OC
76
94
61
94 731 6*2 16* 130
60
47 1*6 6*6 670 3*7
79 446 18* 233 422
96
"
ral5*4 121 17*2*6
42 6*4 14* 22* 1*4
36
37
14 14 14 14 14
643 957 357 357 357
42 6* 12 18 2*
22 1* 61 67 13
79 8 19 23 42
86656
EC/ TC/ ED
TC PM10PM10 PM10 OC EC
76
94
61
0*7 OM 028 94
60
075 014 011 47
0*1 030 024 78
36
0*7 0*2 0*6 Fa"
065 0*4 035 42
96
37
14 14 14 14
357 357 357 643
07 0* 03 42
06 01 01 22
0* 0* 0* 78
670 8*2
6*6 813
432 102
626 163
615 IS*
]121 611
642 150
,
6*8 137
746 171
14 14
643 643
64 12
12 61
7* 17
StaeSamptorB
EC/ EC/
TC OC TC
16* 1>46 0*9
142 1*1 0*4
14*236 070
227 261 072
2022*2 074
733 6*5 0*3
214233 070
193 245 071
246226 069
«O Q
V V
14 14 14
643 643 643
16 2* 07
73 1* 06
25 6 0*
TC.' EC/
PM10PM10
048 029
028 018
031 022
029 021
0*7 042
034 028
051 035
0*4 038
067 047
8Q
V
14 14
643 643
04 03
03 02
07 OS
22 22 22 22 22 22 22 22 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11
%Compl 80* 61* 61* 61* 61* 61* 81* 61* 81* 45* 46* 46* 46* 46* 46* 45* 81* 81* 81* 61* 81* 81* 81* 61*
32 37 10 14 3 07 0* 03
18 04 1* 24 1* 06 01 0*
67 7 19 25 8* 0* 0* 07
43 6 13 18 2* 07 0* 03
22 1* 61 67 1* 0* 01 01
79 8 18 23 42 0* 0* 06
43 62 13 16 27 07 04 03
22 12 61 73 1* 0* 03 02
79 7* 17 26 60* 07 05
WaatondsNact 10 9889889 11111111 11111111
Nai? 10 10 10 10 10 10 10 10 33333333 33333333
UCentpl 100 80* 80* 80* 800 800 BOO 90* 333 333 333 333 933 933 933 933 933 933 933 933 933 333 333 333
8/30-8/30 AD Pairt
Maan
Mn
Max
Nad
29 42 7* 12 22 07 04 03
21 1* 47 62 1* 0* 02 01
40 61 12 20 40* 07 04
2727272727272727
94
94
94
73 9*
73 8*
73 8*
17 13
17 13
17 13
0* 0* 03
0* 0* 03
06 0* 03
94
94
34
67 8*
67 8*
67 86
17
17
17
1* 0*
1* 0*
1* 0*
05 03
0* 03
05 03
N»«p 30 30 90 90 30 X 90 30 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11
UCompI 80* 800 80* 80* 80* 80* 900 80* 45* 45* 46* 46* 46* 46* 46* 46* 61* 81* 81* 61* 81* 61* 61* 81*
Max
91 9* 84 19 27 07 0* 03
16 04 1* 24 1* 0* 01 0*
87 61 IB 26 66 0* 0* 07
46 S* 12 1* 2* 07 0* 03
22 1* 61 67 13 0* 01 01
78 6 IB 23 42 0* 0* 06
42 64 12 16 2* 07 04 03
22 12 61 73 1* 0* 03 02
79 7* 17 26 60* 07 0*
Sto8 126th&Arnata
-------
Table 8. Phase 1 Overall Precision Data from Collocated Samplers
Day No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Day Date
Wed 07/10
Thu 07/11
Fri 07/12
Sat 07/13
Sun 07/14
Mon 07/15
Toe 07/16 •
Wed 07/17
Thu 07/18
Fri 07/19'
Sat 07/20
Sun 07/21
Mon 07/22
Tue 07/23
Wed 07/24
Thu 07/25
Fri 07/26
Sat 07/27
Sun 07/28
Mon 07/29
Tue 07/30
Wed 07/31
Thu 08/01
Fri 08/02
Sat 08/03
Sun 08/04
Mon 08/05
Tue 08/06
Wed 08/07
Thu 08/08
Fri 08/09
Sat 08/10
Sun 08/11
Mon 08/12
Tue 08/13
Wed 08/14
Thu 08/15
Fri 08/16
Sat 08/17
Sun 08/18
Mon 08/19
All days
Site 3 Site 9 Site 10
, 125th & Amsterdam Madison Ave. Canal St.
103 163 di 109 169 di 110 160 dl Phase 1
25.3
30
40.2
15.9
42.8
1
45.7
49 44.6 59.1 28
19.4 20.6 21.9 6.12 14
9.6
35.3
40
53.1 70.9
67.9
36.2 32.7 -10 52.2 47.6 -92 53
37.6 56.3 54.3 -3.6
31.3
55.1 43.9
60.2
57.1 61.2 6.93 64 69.1 7.66 69 52.7 -26.8
»
86.2
88.5 93 90.9 -2.3 93.7
46.5
47.6
23.4 122 119 -2.7 31 31.2 0.6
18.8
42.4
28.6 28.2 -1.4 39 38.4 -1.6 50.5
44
29.7 52.4 55
30.1
49
29
31.9 36 12.1 44.2 71.6
N4 N7 N314N
dj 1.9 dj 3.7 dj -9.9 0.3 D
Sj 9.8 Sj 12.1 Sj 14.8 12.1 Sa
23.9 U95%PL
-23.4 L95%PL
Notes:
percent difference - d . 200'(Yl-Xi)/(XI+Yi)
average analyzer difference - dj - (1/N)*SUM(dl), 1.1 to N
standard deviation - Sj - SQRT((1/{N-1))'(SUM(d!"2)-{1/N)'(SUM(dl))**2)J
average of averages - D - (1/k)*SUM(dj). | - 1 to k
pooled standard deviation -- Sa « SQRT[(1/k)'SUM(Sj**2)]
upper 95% probability limit - U95V.PL « D+1.96'Sa
tower 95% probability limit - LB5%PL - D-1.96*Sa
CH-97-02
32
-------
Table 9. Phase 2 Overall Precision Data from Collocated Samplers
Day No.
1
2
3
4
5
6
7
8
, 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Day
Thu
Frl
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Mon
All days
Date
08/22
08/23
08/24
08/25
08/26
08/27
08/28
08/29
08/30
08/31
09/01
09/02
09/03
09/04
09/05
09/06
09/07
09/08
09/09
09/10
09/11
09/12
09/13
09/14
09/15
09/16
09/17
09/18
09/19
09/20
09/21
09/22
09/23
09/24
09/25
09/26
09/27
09/28
09/29
09/30
Site 6 Site 9 Site 10
207th & 9th Ave. Madison Ave. Canal St.
206 266 d! 109 169 dl 110 160 di Phase 2
60.9 48.6 -4.6 75.9 72.5 -4.6 70.8
56.9
33
26.4 33.6 30.8 -8.7 37.1 37.7 1.6
40.6
40.6
36.8 61.1 54.2 -12.0 47.4
29.8
26.3
23.6 34.1 31.5 -7.9 38
27.2
22.1
38.4 33 -15.1 50.1 49.6 -1.0 58.4 55.1 -5.8
56.7
52
37.3 47.1 46.9 -0.4 75.2
34,2
31.8
40.7
34.1 78.5 75.7 -3.6
22
20 35.5 33.8 -4.9 61.3 59.2 -3.5
17.8
18.4
26.2
26.8
-
13.1 13.4 2.3 | 21.7|
20.1
19.2
42.2 41.4 -1.9 48.6 50.6 4.0
36 35.8 -0.6 49.5 55.8 12.0
36.6 35.1 -4.2 37.9 38.7 2.1
N3 N11 N620N
dj -5.8 dj -4.5 d] 1.7 -2.8 D
Sj 8.8 S] 3.7 S] 6.2 5.3 Sa
7.6 U95%PL
-13.3 L95%PL
Notes:
percent difference - d . 200*(Yi-Xi)/(XI+Yi)
average analyzer difference - dj«(1/N)*SUM(di). 1.1 to N
standard deviation - Sj • SQRT[(1/(N-1)nSUM(dr2M1/N)*(SUM(di)r2)]
average of averages - D - (1/k)*SUM(dj), j. 1 to k
pooled standard deviation - Sa - SORTr.(1/k)*SUM(SjM2}]
upper 95% probability limit - U95%PL • D+1 -96*Sa
tower 95% probability Imit - L95%PL - D-146*Sa
CH-97-02
33
-------
3.4.2 Analytical Precision
For Phases 1 and 2, analytical precision was determined by replicate weighings of clean
•• and exposed filters. Approximately 12.2 percent of clean filters and 13.9 percent of exposed
filters were reweighed. The average mass difference was about 0 pg/filter for both clean and
exposed filters with a standard deviation of the mass difference of about 2 pg/filter (0.3 pg/m3)
for clean filters and 3 pg/filter (0.4 pg/m3) for exposed filters.
For the Phase 2 carbon analysis, analytical precision was determined from the relative
percent difference (RPD) of duplicate measurements.2 Analytical precision for OC was 10.9
percent with a standard deviation of 10.1 percent, for EC it was 5.6 percent with a standard
deviation of 10.9 percent, and for TC it was 7.5 percent with a standard deviation of 10.1
percent. Blank values for the carbon analysis averaged 20 pg/filter (2.8 pg/m3) for OC, 23
pg/filter (3.2 pg/m3) for EC, and 43 pg/filter (6.0 pg/m3) for TC.
3.4J Relative Accuracy
Relative accuracy, the degree of the correctness of data, is determined from the difference
between observed values and values presumed to be true. For this study, relative accuracy of the
- saturation samplers was obtained by comparing their PM10 and PM2.5 values with those of the
PM10 reference dichotomous samplers operated by New York State, which were presumed to be
"true." The results for the reference samplers were obtained from New York State. Tables 10
and 11 show the results for the comparison of saturation samplers with reference samplers at the
two reference sites, Sites 9 and 10, for Phases 1 and 2, respectively. For Site 9, each table gives
the relative accuracy forPMlO and PM2.5 and the ratio of PM2.5/PM10 as determined by the
saturation sampler and the reference sampler. For Site 10, each table gives the relative accuracy
for PM10 and the ratio of PM2.5/PM10 for the reference sampler. In addition, Appendix C gives
a series of plots comparing PM10 and PM2.5 values for the saturation and reference samplers
-a at Sites 9 and 10 for both phases of the study.
For Phase 1, the relative accuracy for PM10 was -2.1 percent with a standard deviation
of 10.7 percent at Site 9 and 3.9 percent with a standard deviation of 22.7 percent at Site 10.
For PM2.5 at Site 9, relative accuracy was 28.7 percent with a standard deviation of 16.9 percent.
The mean ratio for PM2.5/PM10 was about 0.66 for the PM10 reference sampler at Site 9 and
about 0.71 for the one at Site 10. In contrast, the PM2.5/PM10 was about 0.83 for the saturation
samplers at Site 9. The ratio of PM2.5/PM10 as determined by the saturation samplers to that
from the PM10 reference sampler at Site 9 averaged about 1.32.
If only valid (and not suspicious) data pairs are considered, at Site 9 relative accuracy
becomes -1.5 percent for PM10 and 27 percent for PM2.5; the mean ratio for PM2.5/PM10
becomes about 0.61 for the PM10 reference sampler and 0.75 for the saturation sampler, and the
ratio of PM2.5/PM10 as determined by the saturation sampler to that from the PM10 reference
sampler is about 1.28.
For Phase 2, the relative accuracy for PM10 was 3.8 percent with a standard deviation
of 25.9 percent at Site 9 and 12.7 percent with a standard deviation of 26.5 percent at Site 10.
CH-97-02 34
-------
RaMfea Accuracy Raauft* tor Pha*a1
Stta9 • Madison Avanua batwaan 47th A 48th atraaU
PhaMl
Day No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
18
20
21
22
23
24
25
26
27
28
20
30
31
32
33
34
35
36
37
38
39
40
41
AUValuas
Day
Wad
Thu
Fri
Sat
Sun
UMM
IWvn
TIN
Wad
Thu
Fri
Sat
Sun
Mon
Tua
Wad
Thu
Fn
Sat
Sun
Mon
Tua
Wad
Thu
Fri
Sat
Sun
Mon
Tua
Wad
Thu
Fn
Sat
Sun
Mon
Tua
Wad
Thu
Fn
Sat
Sun
Mon
Only Vabd Pah*
Data
07/10
07/11
07/12
07/13
07/14
07/15
07/16
07/17
07/18
07/19
07/20
07/21
07/22
07/23
07/24
07/25
07/26
07/27
07/28
07/29
07/30
07/31
08/01
08/02
08/03
08/04
08/05
08/06
08/07
08/08
08/09
08/10
08/11
08/12
08/13
08/14
08/15
08/16
08/17
08/18
08/19
N
Maan
StdDav
N
Maan
StdDav
Ra( Sat
PM10 PM10
42.7
42.1
54.8 44.6
20 20.6
46.9
70.9
604 522
36 376
59.5 [j||jT]
7o reran
87 930
1479 1223
* 37 390
58.9 524
383 442
14 12
567 580
30 9 27.3
9 9
59.1 562
382 315
«i
FM10
•186
3.0
3.6
-11
1 '7-4
| -86
6.9
-17.3
54
-11.0
154
11
-27
10.7
9
•15
11.6
Rat
PM2.5
25
29
414
105
367
37.4
24.9
43.8 1
586|
691
303
259
33
275
14
352
14.8
8
323
168
Sat
PM2.5
16
57
44
26
rm
nrn
79
41
31
45
37
11
466
195
8
399
185
dj
PM25
53.3
184
6.0
| 55.0
| 160
13.9
36.6
17.8
367
331
10
28.7
16.9
8
27.0
155
Rat
25/10
0585
0.689
0.755
0525
0783
0.742
0.655
0.736
0.837
0.794
0205
0700
0560
0718
14
0663
0.16
6
0.612
0.19
Sat
25/10
0782
0549
0.702
1232
1.063
0.846
0.339
0.782
0861
0.828
10
0528
023
8
0.748
017
Sat/Rat
25/10
1.489
1.144
1.072
1.674
1269
1.065
1652,
1.117
1536
1.153
10
1.317
024
8
1279
024
SftalO
Rat
FM10
364
475
46.1
19.5
64.8
SS2
51.6
622
646
85.3
28.3
451
538
47.3
14
50.6
164
10
51.3
185
- Canal Straat and Broadway
Sat
PM10
14.0
530
56.3
43.9
690
937
310
505
550
716
10
53.8
220
10
53.8
22.0
«i
PM10
•282
-40
9.1
•294
6.8
98
95
12.0
22
514
10
3.9
22.7
10
3.9
227
Rat Rat
PM25 25/10
23.8 0654
33.1 0692
345 0748
9.6 0492
469 0.724
40 0.725
36.7 0.711
466 0749
528 0.817
68 0797
195 0689
344 0763
366 0.680
308 0651
14 14
367 0707
144 008
10 10
375 0708
164 009
Notas
paroantaga dWaranoa - dj. 100*(Sat-Raf)/Raf
Sat/Rat 2.5/10 «(Saturation Samplar PM2 5/PM10 Ratk>V(Rataranca Samplar PM25/PM10 Rate)
SuaptoiousData
CH-97-02
35
-------
Tab* 11. Ralativa Accuracy Raaufta tor Phaaa 2
8Ha9 -MadwonAvanuabatwaan 47th* 48th atraats
*haaa2
Day No
1
2
j 3
4
5
6
7
a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
£ 40
,* AlVahMS
Day
Thu
Fri
Sat
Sun
Mon
TIM
Wad
Thu
Fn
Sat
Sun
Mon
TIM
Wad
Thu
Fri
Sat
Sun
»j__
Mon
TIM
Waa
Thu
Fri
Sat
Sun
fcj_.»
won
TIM
Wad
Thu
Fn
Sat
Sun
Mon
TIM
Wad
Thu
Fri
Sat
Sun
Mon
Data
0802
0803
0804
08O5
08/26
08/27
08/28
08/29
08/30
08/31
09/01
09/02
Oft/03
09/04
09/05
09/06
09/07
09/08
Oft/09
09/10
69/11
09/12
09/13
OB/14
09/15
09/16
09/17
09/18 ,
09/19
09720
0901
0902
09O3
09/24
09/25
09/26
09/27
0908
0909
09/30
N
Maan
StdDav.
Raf
PM10
72.6
285
51
29
464
42
361
467
6
440
142
Sat dj
PM10 PM10
75.9 45
33.6 17.9
61.1 19.8
34.1 175
60.1 8.0
472 124
785
355
1 2171-535
' 422
36.0
366
12 7
460 38
17.6 25 9
Raf
PM25
50.9
173
354
15.4
284
26
202
35.1 1
,
8
286
11.7
Sat
PM25
61.1
27.1
453
22.9
404
37.6
597
286
156
273
31.9
12
346
14.9
dj Raf Sat
PM25 25/10 25/10
20.0 0.701 0.805
566 0.607 0.807
26.0 0.694 0.741
487 0531 0.672
423 0.612s 0.806
446 0.619 0.797
0560
•49.9 0.752 0.811
787
272 0634 0.777
362 0.08 0.05
SafRaf
25/10
1.148
1429
1.068
1265
1.317
1267
1.079
7
121
0.11
SHalO
Raf
PM10
67.3
33.8
63.7
24.3
485
66.4
106
294
30.1
\
9
41.8
21.0
- Canal Straat and Broadway
Sat dj
PM10 PM10
70.8 62
37.1 9.8
47.4 -256
38 564
58.4 20.4
752 99
61.3
486
49.5
379
10 6
524 127
136 265
Raf
PM25
491
227
484
17
328
47.6
0.6
205
18.8
9
286
17.0
Raf
25/10
0.730
0672
0.760
0.700
0676
0696
0.057
0697
0.625
9
0623
022
Only Valid Pair* N 666666666
Maan 449 503 13.4 289 391 40.0 0.627 0.771 124
StdDav. 164 163 6.1 13.1 137 13.6 0.06 0.05 010
66666
51.0 545 12.7 363 0705
187 163 265 142 0.03
Notes
paroantaga dWaranea - d) • 100*(Sat-RatyRaf
Sat/Raf2 5/JO • (Saturation Samptar PM2.5/PM10 RattoWRafaranoa Samplar PM25/PM10 Ratio)
CH-97-02
36
-------
For PM2.S at Site 9, relative accuracy was 27.2 percent with a standard deviation of 36.2 percent.
The mean ratio for PM2.5/PM10 was about 0.63 for the reference sampler at Site 9 and about
0.62 for the one at Site 10. In contrast, the PM2.5/PM10 was about 0.78 for the saturation
samplers at Site 9. The ratio of PM2.S/PM10 as determined by the saturation samplers to that
from the reference sampler at Site 9 averaged about 1.21.
If only valid (and not suspicious) data pairs are considered, at Site 9 relative accuracy
becomes 13.4 percent for PM10 and 40 percent for PM2.5; the mean ratio for PM2.5/PM10 stays
about 0.63 for the PM10 reference sampler and 0.77 for the saturation sampler; and the ratio of
PM2.5/PM10 as determined by the saturation sampler to that from the reference sampler is about
1.24.
These results suggest that the relative accuracy for PM10 is fairly good overall, although
the results for individual measurements may vary considerably. However, the results for PM2.S
suggest that the results for the saturation sampler are biased high by about 28 percent compared
to the reference dichotomous sampler.
For the Phase 2 carbon analysis, laboratory relative accuracy was determined from
analytical recovery data for standards and splits.2 For standards, the analytical recovery was
about 93.6 percent with a standard deviation of about 5.4 percent; for splits, the analytical
recovery for percent OC was about 100.7 percent with a standard deviation of about 8.9 percent
3.5 Other Data
Some data used for the data analysis in this report were obtained from other sources. As
mentioned in Section 3.4, reference sampler data were obtained from New York State. In
addition, meteorological data were obtained from the weather report in the New York Times. Due
to limited resources, these meteorological data were used primarily to assess unusual conditions
that might significantly impact paniculate matter levels. Such conditions might include a frontal
passage that results in low weekday concentrations at all sites, significant amounts of rain (> 0.25
inches) that could reduce particulate levels, and periods of stagnation in which paniculate levels
could be built up to high levels and remain high at all sites. In addition, data on daytime PM2.5
concentrations and traffic counts (automobiles, buses, trucks, and pedestrians) were obtained from
Professor Patrick Kinney of the Columbia School of Public Health, who was conducting a
parallel study.5 Table 12 gives the hourly average traffic counts for four Phase 1 Harlem
Community Sites (Sites 1,3,5, and 6). The hourly averages were taken over the period 10 a.m.
to 6 p.m. and represent averages taken over two or three days. Also, the results of a 1993 study
of the impact of mobile source emissions on ambient particulate levels at the midtown Manhattan
reference site were used for comparison to the results obtained in this study.6
CH-97-02 37
-------
Table 12. Average Traffic Counts for Select Harlem Community Sites
,
Type of Traffic
Autos
Trucks
Buses '
Pedestrians
£
i
CD
I
00
•o
§
1
288
9
4
55
1
te
Ki
i
00
£
U»
CM
3
2297
175
133
1406
g
c
I
00
£
§
5
1791
66
50
2095
CO
™Q
i
0
CD
T?
LU
6
153
6
2
141
Notes:
i i
Data obtained from Prof. Patrick Kinney of Columbia University
Average traffic over 10 a.m. to 6 p.m. time period for 2 or 3 days
CH-97-02
38
-------
4.0 DATA ANALYSIS
This section builds upon the results reported in Section 3 and discusses additional
summary statistics and analyses used to try to determine the answers to the two questions given
in Section 1.
4.1 PM10 Summary Statistics for Each Phase by Site
Tables 1 and 2 summarize the PM10 data for Phases 1 and 2, respectively. As discussed
in Section 3.1, Tables 1 and 2 show considerable variation in PM10 concentrations from site-to-
site on a given day, and from day to day for a given site.
4.1.1 Phase 1
During Phase 1, mean concentrations for the sites varied from a low of 28.3 pg/m3 at Site
6 (the control site) to a high within the Harlem community of 40.5 pg/m3 at Site 3 and a high
of 58 ug/m3 at the midtown Manhattan reference site. Table 1 shows that mean concentrations
are, in general, higher on weekdays than on weekends when traffic is expected to be less. Site
9 is the exception since its results are skewed by the data point for Saturday, August 10, 1996,
when construction at the site resulted in abnormally high PM10 concentrations.
As discussed in Section 2.3, equipment and logistical problems resulted in substantially
poorer completeness than desired or expected. Table 1 shows that completeness during Phase
1 varied from day to day with a low of 37.5 percent and a high of 100 percent. Over all of
Phase 1, completeness for each site varied from a low of 30 percent for Site 4 to a high of about
86 percent for Sites 9. The expected data completeness for each site was 80 percent; only three
sites, Sites 1, 7, and 9, met this criterion. In a number of respects, the poor data completeness
at most sites made data analysis more difficult and will be discussed in greater detail below.
Table 1 also shows that sites could exhibit significant variation in PM10 concentrations
on a given day and a given site could exhibit significant variation in PM10 concentration on
different days. Excluding exceptional and suspicious data, on August 19, with a mean PM10
concentration of about 33 pg/m3, PM10 concentrations varied from a low of 22 pg/m3 at Site 6
to a high of 72 pg/m3 at Site 10, the downtown reference site. On the other hand, on July 19,
with a mean concentration of about 23 pg/m3, PM10 concentrations varied from a low of 21
pg/m3 at Site 6 to a high of only 25 pg/mj at Site 5. Table 13 shows the same data as Table 1,
but it reports the standard deviation and coefficient of variation (CV) of the data instead of
minimum and maximum PM10 concentrations. The standard deviation is a measure of the root
mean square deviation of values from the mean, and the CV is simply the standard deviation
divided by the mean and expressed as a percent Table 13 shows that from day to day, the CV
varied from a low of 4.5 percent to a high of 90 percent If days with exceptional values and
suspicious data points are not considered, then CVs varied from a low of about 4.5 percent to
a high of 42.3 percent. In general, the highest CVs occurred on days when the reference site
samplers were sampling and the lowest CVs occurred during days with rain, frontal passages, or
stagnation.
CH-97-02 39
-------
Table 13 VwlabWiy in PMIODatt tor Pha*el-Hartem Community
Day No
i
2
3
4
6
e
7
•
«
to
11
12
13
14
16
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
> Day
Wed
Thu
Fn
Sai
Sun
Mon
Tue
Wad
Thu
Fri
Sat
Sun
Mon
Tue
Wad
Thu
Fn
Sal
Sun
Mon
Tue
Wad
Thu
Fn
Sai
Sun
Mon
Tue
Wed
Thu
Fn
Sal
Sun
Mon
Tue
Wad
Thu
Fn
S*
Sun
Mon
Date
07/10
07/11
07/12
07/13
07/14
07/15
07/16
07/17
07/18
07/19
07/20
07/21
07/22
07/23
07/24
07/25
07/26
07/27
07/26
07/29
07/30
07/31
08/01
08/02
08/03
08/04
08/05
08/06
06/07
0808
08/09
08/10
06/11
08/12
08/13
08/14
08/15
08/16
OB/17
OB/18
08/19
1
18
33
17
33
41
39
39
48
22
12
66
33
60
34
11
22
37
26
27
51
69
78
77
79
38
n§~
»
15
14
28
40
34
23
27
26
|
2
29
39
13
26
38
81
83
40
] 4*
33
25
24
20
44
38
27
47
29
i
as
30
40
16
43
46
49
\ 19
10
35
40
63
68
36
31
60
,57
86
89
47
48
23
18
42
29
44
30
30
49
29
32
i
13
22
16
43
29
33
22
i
31
42
38
26
15
17
35
20
16
28
S3
64
62
70
36
19
14
26
36
31
25
26
24
23
i
24
17
44
42
47
25
11
33
47
51
34
17
20
35
29
18
36
63
62
76
76
78
41
60
18
18
34
48
26
i
16
34
22
29
38
21
12
cn
41
25
18
17
30
18
16
o~
T6~
54
47
64
69
34
34
21
13
30
30
22
20
25
22
I
25
31
20
38
35
36
44
24
11
26
J
43
55
18
20
35
22
22
1 39
54
64
65
88
82
41
44
16
27
23
36
36
23
47
22
27
«
|
23
49
.42
7
59
44
68
43
47
12
12
29
43
39
27
67
32
32
I
45
21
71
62
38
nsn
CE~
93
122
39
52
44
rt
I
10
14
63
66
1 **
1 69
94
31
51
55
72
Nad
6
4
3
6
4
6
6
6
6
5
6
4
4
7
7
4
7
6
S
6
7
7
6
6
6
, 6
3
7
8
7
6
7
8
6
8
7
10
7
7
6
10
Nexp
10
8
8
8
10
8
6
10
8
8
10
8
8
10
8
8
10
8
8
S
10
6
10
8
8
10
8
8
10
8
8
10
8
8
10
8
10
8
6
6
10
%Compl
600
600
375
625
400
625
625
600
750
625
600
600
600
700
875
600
700
625
625
625
700
876
600
750
625
600
376
876
800
875
1000
700
1000
750
800
875
1000
875
875
750
1000
Mean £
189
264
347
172
370
369
341
406
442
226
155
11 1
.363
313
496
630
371
159
190
343
299
193
342
645
610
687
746
803
832
403
432
384
164
243
296
390
379
253 -
41 9
264
333
ild Dev
46
29
40
14
38
81
71
64
41
14
39
09
128
125
96
109
106
25
20
23
124
76
162
37
60
64
34
79
77
36
92
346
39
96
98
48
64
28
125
32
141
CV
241
11 5
115
61
103
220
206
134
92
69
250
80
357
399
193
205
286
157
106
67
414
389
473
66
9B
142
45
99
93
69
21 3
801
236
405
332
124
222
109
29B
121
423
MDays Nad 35 IB 31 31 29 31 34 18 12 10 I ISutptoout data
Nexp 41 4T 41 41 41 41 41 41 14 14
%CornplB64 439 756 766 707 766 829 439 857 714
Mean 357 380 405 303 390 283 364 378 660 538
Sid Dev 183 18 16 14 19 157 179 188 261 209
CV 812 474 445 461 467 653 491 497 45 368
Weekdays Nad 26 14 22 22 21 21 24 14 9 7
Nexp 2929292929292929 10 10
%Cornpl697 483 769 769 724 724 828 483 900 70.0
Meen 393 406 460 316 417 294 383 384 643 606
Std Oev 173 166 17 124 177 164 17 194 166 16.6
CV 44 46 378 394 425 666 443 604 306 258
Weeliand* Nad 9 4 9 9 8 10 10 4 3 3
Nexp 12 12 12 12 12 12 12 12 4 4
«Compl760 333 760 760 667 033 833 333 760 760
Mean 262 294 294 274 321 261 319 369 690 380
Std Dev 169 123 166 168 205 137 192 166 417 23
CV 673 42 625 616 638 527 603 463 604 605
Parser* Complete (% Compf) - lOO*(Acrual Number (NactJ/Expedad Number (Nexp]}
Coefloerd ot Variation (CV). 100*(Std Dev /Mean)
CH-9742
40
-------
Table 13 also shows the variability of data at each site for Phase 1. In general, the
variability at each site over the period of Phase 1 is greater than the variability among sites on
a given day. During Phase 1, CVs ranged from a low of 38.8 percent at Site 10 to a high of 55.3
percent at Site 6, the control site. In general, sites with the highest mean PM10 concentrations
have the lowest CVs. In general, the CVs for weekday sampling are greater than those for
weekend sampling. However, Sites 2, 6, and 8 are exceptions to this observation.
The fact that the variability of data from day to day at a given site is usually greater than
the variability in data from sites on a given day when coupled with the poor data completeness,
for some sites, makes it difficult to answer the question of which site(s) have the highest mean
concentrations. As discussed in Section 3.1, one way to address this difficulty is to use relative
rankings. Table 3 shows that if the mean site ranking is used as the indicator, then the site with
the highest PM10 concentrations, other things being equal, should be Site 10 (the downtown
reference site) followed by Site 9 (the midtown reference site) and Site 3 (the Harlem community
site at 125th Street and Amsterdam Avenue). Figure 11 shows mean PM10 concentrations by
type of site. Site 9, the midtown Manhattan street canyon reference site, and Site 10, the
downtown Manhattan reference traffic site, look like they might violate the current PM10 annual
average of 50 pg/m3,7 In addition, counts of daytime (10 am to 6 pm) automobile, bus, truck,
and pedestrian traffic at four sites (Sites 1, 3, 5, and 6) were obtained from Professor Kinney at
Columbia University and are given in Table 12. The counts were averages for several weekdays
in July 1996. Average hourly automobile traffic varied from a low of 153 to a high of 2300;
hourly truck traffic varied from a low of 6 to a high of 175; hourly bus traffic varied from a low
of 2 to a high of 133; and hourly pedestrian traffic varied from a low of 55 to a high of 2095.
Except for pedestrian traffic, all lows occurred at Site 6, the control site, and all highs at Site 3;
for pedestrian traffic, the low occurred at Site 1, a bus depot site, and the high at Site 5, the
Harlem Hospital site. For the period examined, the bus depot site, Site 1, only averaged 4 buses
an hour.5 In April 1993, the Desert Research Institute had the following hourly average traffic
counts for the midtown Manhattan reference site, Site 9: 15 motorcycles, 1197 cars, 63 trucks,
and 171 buses.6
4.1.2 Phase 2
During Phase 2, mean concentrations for the sites varied from a low of 21.9 pg/m3 at Site
5 to a high within the Washington Heights community of 35.8 pg/m3 at Site 7 and a high of 52.4
pg/m3 at the downtown Manhattan reference site. Table 2 shows that mean concentrations are,
in general, higher on weekdays than on weekends when traffic is expected to be less. Site 5 is
the exception.
Table 2 shows that completeness during Phase 2 varied from day to day with a low of
0.0 percent and a-high of 100 percent. Over all of Phase 2, completeness for each site varied
from a low of 20 percent for Site 5 to a high of about 93 percent for Sites 6 and 8. The
expected data completeness for each site was 80 percent; five sites, Sites 1, 3, 6, 8, and 9, met
this criterion.
Table 2 also shows that sites could exhibit significant variation in PM10 concentrations
on a given day, and a given site could exhibit significant variation in PM10 concentration on
CH-97-02 41
-------
9
$
6
«.• l»
I-
80
CO*
I
feo
.o
I
I 40
o
o>
&
-------
different days. On September 12, with a mean PM10 concentration of about 32.5 pg/m3, PM10
concentrations varied from a low of 19.7 pg/m3 at Site 1 to a high of 61 pg/m3 at Site 10, the
downtown reference site. On the other hand, on September 5, with a mean concentration of
about 54.8 pg/m3, PM10 concentrations varied from a low of 52 pg/m3 at Site 6 to a high of 56.9
pg/m3 at Site 8. Table 13 shows the same data as Table 2, but it reports the standard deviation
and coefficient of variation (CV) of the data instead of minimum and maximum PM10
concentrations. Table 14 shows that from day to day the CV varied from a low of 3.3 percent
to a high of 43.4 percent
Table 14 also shows the variability of data at each site for Phase 2. In general, the
„ variability at each site over the period of Phase 2 is greater than the variability among sites on
, a given day. During Phase 2, CVs ranged from a low of 21.7 percent at Site 5 to a high of 40
percent at Site 3. In general, the CVs for weekday sampling are greater than those for weekend
sampling.
Table 4 shows that if the mean site ranking is used as the indicator, then the site with the
highest PM10 concentrations, other things being equal, should be Site 10 (the downtown
reference site) followed by Site 9 (the midtown reference site) and then Site 2 (the Washington
Heights community site at 179th Street and Broadway) and Site 8 (the Harlem community site
with the highest ranking in Phase 1). Figure 12 shows mean PM10 concentrations by type of
, site. Only Site 10, the downtown Manhattan reference traffic site looks like it might violate the
fc PM10 annual average.
4.2 Comparison of PM10 Concentrations Among Sites
I For both Phases 1 and 2, the two reference sites, Site 9 in midtown Manhattan and Site
10 in downtown Manhattan, tend to have higher PM10 concentrations than any of the sites in
either the Harlem or Washington Heights communities. The question is whether this difference
T in PM10 concentrations is significant Two tests were used to try to assess whether the
differences in PM10 concentrations between a given community site and a reference site were
significant: the Student's t-test8 and the Wilcoxon signed rank test9'10 For each phase, each
comparison used only days when both sites being compared had PM10 values.
The Student's t-test calculates a statistic, t which is a function of the number of pairs,
their mean values, and standard deviations. If t is greater than a certain critical value ^ then the
difference in mean values is significant at a certain probability. For these tests, values of t,. were
chosen to correspond to a 95 percent probability that the two means were different The
Wilcoxon signed rank test is a nonparametric test In this test the differences between two
values is calculated and each difference is assigned a rank with 1 for the smallest difference and
N for the largest difference if there are N non-zero difference pairs. The sum of all negative
differences and the sum of all positive differences are then calculated. The statistic, T, is the
smaller of the two sums, and is then compared to a critical value, Tc, based on the total number
of pairs and the probability desired of being certain a difference is significant If T is less than
Tc, then the median of one set of data is significantly different from the other. The two tests may
give different results. If there is relatively large variability among the daily values at each of two
sites, but on most days one site has consistently higher (or lower) values than the other site, then
CH-97-02 43
-------
Tibto 14 Vatiato% In PMlODaU tor Phaja 2-Wa»NngU)nHa$h» Community
Mill'
«i«;;*..-s
I i I I I I ! I ! 3
Dav No
1
2
3
4
6
6
7
8
9
10
11
12
13
14
15
16
17
•18
19~
20
21
22
23
24
25
26
. 27
28
29
30
31
32
33
34
36
36
37
38
39
40
Day
Thu
Fri
8«
Sun
Mon
Tue
Wad
Thu
Fri
Sat
Sun
Mon
Tue
Wed
Thu
Fn
Sai
Sun
Mon*
Tue
Wad
Thu
Fn
Sal
Sun
Mon
Tue
Wad
Thu
Fn
Sai
Sun
Mon
Tue
Wad
Thu
Fn
Sai
Sun
Mon
Data
0822
0823
0824
0825
0826
0627
0828
0829
OS/30
0801
OB/01
09/02
09/03
09/04
09/05
08/06
06/07
06/08
09/09
09/10
09/11
09/12
09/13
09/14
09/15
09/16
09/17
09/18
09/19
0920
0921
0922
0923
0924
0925
0926
0927
0928
0929
09/30
1
37
31
42
43
43
30
28
27
43
64
64
35
37
35
39
47
24
20
17
t7
25
31
22
26
24
2
34,
SO
68
36
87
48
21
24
SB
25
27
37
33
29
35
17
23
27
23
31
49
60
27
37
48
36
35
27
28
28
40
64
67
33
32
31
37
44
24
23
19
14
22
60
14
22
15
66
31
40
32
26
22
32
31
16
22
12
21
30
18
25
16
16
22
27
21
61
67'
S3
26
41
41
37
SO
26
24
27
22
SB
67
62
34
32
41
34
22
20
18
18
26
27
13
20
19
63
66
35
27
43
47
SB
34
29
26
38
64
33
43
22
47
18
18
23
23
61
68
34
29
44
44
40
29
23
28
25
42
67
49
40
39
44
60
33
32
31
21
27
33
20
25
29
29
33
24
33
25
22
19
37
26
27
76
34
61 -
34
60
47
79
36
ran
42
36
37
10
708
371
474
SB
664
752
613
486
495
379
Nan
7
4
6
6
6
4
6
6
4
7
6
6
7
4
5
7
6
6
6
6
6
6
6
7
6
6
0
7
6
6
2
3
3
4
3
3
4
3
3
6
New
10
6
8
10
8
6
10
8
6
10
8
8
10
6
6
10
8
8
6
10
8
10
8
8
10
8
6
10'
8
10
3
3
3
6
3
3
6
3
3
6
%Comtt
700
600
625
600
626
600
600
750
600
700
626
750
700
600
625
700
625
750
625
600
626
BOO
625
875
600
625
00
700
750
600
667
1000
1000
600
1000
1000
800
1000
1000
1000
Mean
694
627
340
306
408
443
438
366
318
290
276
269
444
630
648
439
362
341
407
601
250
326
206
183
241
357
163
229
228
329
v 316
234
372
256
166
320
284
254
308
Sid Dav
97
46
19
37
23
30
88
45
67
61
10
30
6B
40
18
143
27
31
^Z3"
137
41
141
62
32
20
79
47
20
53
36
14
44
90
77
24
116
64
19
62
CV
163
72
65
122
56
69
200
127
210
176
38
116
163
63
33
325
78
•92
66
273
166
434
252
173
82
221
256
87
232
109
43
189
241
300
130
370
226
75
202
MOiyt Naet 25 20 27 12 8 28 20 37 12 10
Naxp 3040303040303040 14 14
«Compl833 600 900 400 200 833 667 925 857 714
Maan 337 321 339 285 219 316 368 343 460 624
Sid Dav 114 107 136 111 47 118 127 115 168 129
CV 339 333 400 368 817 378 366 335 365 247
I ISmoctom date
WMkdaya Nad 18 13 21 6 6 21 16 25 10 ,8
Naxp 2228222228222228 11 11
*ComplB18 464 955 364 179 965 682 893 909 727
Mean 364 329 367 307 193 S32 396 362 486 661
8»d Dav 128 126 143 132 34 133 130 130 174 116
CV 362 861 361 429 174 899 329 868 360 211
Waakanft
Nao 6776366 12 22
Naxp 6 12 12 12 12 12 12 12 3 3
«CompttOO 663 683 417 260 667 600 100 667 667
SOD 306 269 263 262 276 269 304 S39 376
Std Dav 63 66 60 66 34 46 66 69 03 06
CV 210 161 230 223 130 179 206 193 07 12
Parcant CompMa (% OompQ • lOO*(Aaual Numbar (Nacq/Expaaad Numbar (Naxp])
Ocwncwnt of Variation (CV)« 10tT(SU Dav /Maan)
CH-97-02
44
-------
- (p. ,
Manhattan Community Based Particulate Study
Mean PM10 Concentrations by Site Type for Phase 2 - Washington Heights Community!
60
CO
40 -
1
-------
the Student's t-test is not likely to show a significant difference while the Wilcoxon signed rank
test might On the other hand, the Wilcoxon test is not very sensitive when there are only a few
paired values and cannot be used at all if the number of pairs is less than six. Thus, when there
are low numbers of pairs, the Student's t-test may show a significant difference when the
Wilcoxon signed test would not or could not be used.
Table 15 shows' the results of assessing the differences in PM10 concentrations using
these two tests. The table shows the results for each pair of sites for each phase of the study.
For both phases, Sites 9 and 10 in general have significantly higher concentrations than the
Harlem community sites; instances in which they do not are primarily due to having too few data
pairs for the tests to be done or to be reliable. For Phase 1, Site 4 (the control site) and Site 6
in general have significantly lower PM10 concentrations than most other sites. These results are
in general agreement with previous observations based on Tables 1 and 3. For Phase 2, Site 6
(the control site) in general has significantly lower PM10 concentrations than most other sites
while Site 8 (the Phase 1 Site 3 Harlem community site) in general has significantly higher PM10
concentrations than the other community sites in Washington Heights. These results are in
genera] agreement with previous observations based on Tables 2 and 4.
4.3 Comparison of Phase 1 and Phase 2 PM10 Levels
Tables 1, 2,13, and 14 show that PM10 concentration levels were generally lower during
Phase 2 compared to Phase 1. At Site 3 in Phase 1 (or Site 8 in Phase 2), PM10 concentrations
'averaged about 40.5 ±18pg/m3 while for Phase 2 they were 34.3 ±11.5pg/m3 (lower by about 15
percent). For the midtown Manhattan reference site, Site 9, the levels were 58 ±26.1 pg/m3 for
Phase 1 and 46 ±16.8pg/m3 for Phase 2, about a 21 percent decrease with the exceptional event
during Phase 1 and about a 12 percent decrease if it was excluded. For the downtown reference
site, Site 10,,the PM10 concentrations were 53.8 i2d.9pg/m3 for Phase 1 and 52.4 ±12.9pg/m3
for Phase 2, about a 3 percent decrease. The decrease observed at Site 3/Site 8 is likely to be
more correct since the number of PM10 concentrations obtained during each phase (31 in Phase
1 and 37 in Phase 2) was much greater at this site than the other two sites (12 for Site 9 and 10
for Site 10 during Phases 1 and 2).
i
4.4 PM2.5 Concentrations and the Proposed PM2.5 Standards
As discussed in Section 3.2, PM2.5 samplers were located at Sites 1, 5 and 9 for Phase
1 and at Sites 1, 2, and 9 for Phase 2. During Phase 1, Table 5 shows that mean PM2.5
concentrations varied from 28 pg/m3 at Site 5 to 47 pg/m3 at Site 9, while during Phase 2, Table
6 shows that they varied from 21 pg/m3 at Site 2 to 35 pg/m3 at Site 9. The proposed annual
standard for PM2.5 is 15 pg/m3." All of the sites that sampled PM2.5 during both phases of this
study would not meet the proposed standard if the mean concentrations observed were typical
of those throughout the year. The proposed 24-hour PM2.5 standard is 50 pg/m3.11 During
Phase 1, there were four exceedances of die proposed standard observed at each of the three sites
while during Phase 2, there were two exceedances observed at Site 9.
The mean ratio of PM2.5/PM10 observed over all sites for both phases was 0.78. This
average ratio is in agreement with values obtained earlier in Philadelphia in a 1994 study using
CH-97-02 46
-------
Table IS. Significance of Site Concenlratnn Differences for PhaMi 1 and 2
Phaw t
ill'
•*•»•*'
ill.
6
ii • i ' i
! « 1 fi « i «
! ! I I i I I
She Test 123456789 10 Notes
1 S
W
2 S no
W no
3 S no no
W yM no
4 S no yM yM
— »••• . —
yM y** ye*
S • two-tailed Student's Mast significant at O.OS level
W • Wilcoxon signed rank test significant at O.OS level
• • number of pairs <-6
** • too few points for test
S S no no no no
W no no no yi
M 1
6 S no no yM no VM
— ..__ ••*• ...
yM y y *
»* yM
7 'S no no no no no no
W no no yes yi
w no yM
B S no . no no no no no no
W yM no* no no no yM no
9 S yM no no y<
W yM yes* no yi
10 S yM no* no y<
W— .-,., MM*> *• BM
jrV* no y
Phase 2
A •
ffl 1 C! i
III!
M no yee no no
is ^f99 y^*^ y^*^
M yM yM no yM no
ra yes yM yM ** **
* S
i - I 1 1 | a •
i i I I i 1 i
Site Test 1 2 3 4 5 6 7 6 9 10 Notes
1 S
W
2 S no
W ym
3 S no no
W yM yM
4 S no no* no
W yM ** no
6 S " no
W " no
•
*
6 S no no no no **
W yM yM no no **
7 S no no* no no ** no
W no no no - yM
6 S no no no no yM no no
W no no yM yi
M yM yM yM
9 S no no* yee no* yea* yM no* no
W yM no* yM
< y^^ y
10 S yM* yM* yM no* no* yM no* yM no
W " ~ yM
" yM' " yM no
CH-97-02
47
-------
saturation samplers.12'13 As discussed in Section 3.4, PM2.5 concentrations measured by the
saturation samplers appear to be biased high when compared to PM2.5 concentrations obtained
from the PM10 reference dichotomous sampler used at Site 9; however, the dichotomous sampler
is not a reference or equivalent sampler for PM2.5. The PM2.5/PM10 ratio deter.nined from the
dichotomous samplers appears to average about 0.66. Using the reference sampler PM2.5 data,
the mean PM2.S concentration was 35 ug/m3 at Site 9 and-37 ug/m3 at Site 10 during Phase 1.
For Phase 2, the mean PM2.5 concentrations were 29 Mg/m3 and 29 ug/m3 at Sites 9 and 10,
respectively. All of these mean concentrations exceed the proposed annual PM2.5 standard. In
addition, during the course of both phases, there were three exceedances of the proposed 24-hour
PM2.5 standard observed at Site 9 and one at Site 10.
4.5 Carbon Analysis
As discussed in Section 3.3, EC was about 30 percent of PM10 while OC was about 15
percent of PM10. Elemental carbon or soot often comes from the exhaust of diesel vehicles,
especially trucks and buses. Available resources did not permit a detailed source apportionment.
•j ,
The two Phase 2 sites chosen for sampling carbon, Sites 8 and 9, were chosen because
they were expected to have high diesel bus and truck traffic. The fraction of EC/PM10 was
similar at both sites, 0.32 at Site 8 (Phase 1, Site 3) and 0.31 at Site 9, but Site 9 generally had
higher PM10 concentrations. The fraction of EC/PM10 would not be expected to be as high at
sites with less diesel bus and truck traffic. In addition, diesel buses and trucks also contribute
to OC.
However, the observed fractions of EC/OC/TC and their percentages of PM10 are
consistent with other studies at sites with heavy diesel bus and truck traffic. Desert Research
Institute did perform a detailed source apportionment from data obtained at the midtown
^Manhattan site (Site 9), which might have higher diesel bus traffic than community sites, in
*1993.6 Using assumed profiles and a variety of paniculate measurements and the Chemical Mass
Balance Model (CMB) version 7, they found that the diesel contribution varied from between 31
and 68 percent of the total paniculate loadings and averaged about 53 percent of ambient PM10
mass; no estimate was given of the relative contribution of trucks and buses. Automobiles were
found to contribute between 0 and 21 percent with an average of about 6 percent. The total
mobile source related contribution (diesel, automobile, and road dust) was responsible for about
two-thirds of PM10 mass. Sea salt contributed about 6 percent and transport-related components,
such as ammonium sulfate and nitrate, accounted for about 22 percent of the mass.
4.6 Unusual Events and Weather
Local weather reports were acquired to determine special weather conditions which might
influence paniculate matter concentrations. Weather comments were included in Tables 1 and
2 to clarify why some peculiar PM10 concentrations were observed. High PM10 concentrations
on the weekend of August 3 and 4 were associated with a week-long stagnation period and low
concentrations on some days were associated with periods of rain or a frontal passage.
CH-97-02 48
-------
4.7 QA/QC Results
Data quality indicators for precision and relative accuracy were discussed in Section 3.4.
This section discusses how well the study met the data quality objectives (DQOs) given in the
Quality Assurance Project Plan (QAPP).2
4.7.1 Completeness
The QAPP goal for daily completeness was 80 percent of sites for a given day. During
Phase 1, that goal was met only about 37 percent of the time; the average daily completeness was
about 70 percent During Phase 2, the daily completeness goal was met only about 30 percent
of the time; the average daily completeness was about 70 percent As discussed in Section 2.3,
operational and equipment problems prevented this goal from being reached.
The QAPP goal for completeness for a site for the study was 75 percent During Phase
1, 7 of 10 sites met the goal; average site completeness was 71 percent During Phase 2, only
half the sites met the goal; average site completeness was 69 percent .The QAPP goal of 80
percent completeness for the entire study period was not met; the average completeness was
about 70 percent
4.7.2 Precision
The QAPP goal for daily overall sampling precision for saturation samplers was ±20
percent. During Phase 1, 2 out of 14 pairs did not meet the goal, but during Phase 2, all 20
sample pairs met the goal. The QAPP goal for overall sampling precision was ±15 percent for
each phase and the study. That goal was met for both phases and the study. For Phase 1, it was
0.3 percent; for Phase 2, it was 2.8 percent; and for the study, it was -1.5 percent For the
carbon analysis filters, these goals for sampling precision were the same. For the carbon filters,
4 of 5 duplicate pairs exceed the 20 percent goal for OC, none exceeded it for EC, and 1
exceeded it for TC. Overall, the average sampling precision was 11.6, 1.6, and 4.6 for OC, EC,
and TC, respectively.
The QAPP goal for gravimetric filter reweighings was the greater of ±20pg/filter or ±15
percent None of the 122 exposed filters that were reweighed exceeded these criteria. In
addition, all the filters met the criterion of an average percent difference within ±10 percent
The goal for analytical precision for the carbon analysis was ±20 percent for OC/EC/TC
from individual analyses of duplicate plugs and ±15 percent for all duplicates. The mean
analytical precision was 11, 6, and 7.5 percent for OC, EC, and TC, respectively. Two out of
five duplicate analyses for OC did not meet the 20 percent criterion; the two values were 21 and
22 percent
4.7J Relative Accuracy
The QAPP goals for relative accuracy were ±20 percent for collocated reference sampler
and saturation samplers for daily differences and ±15 percent for the average percent difference
CH-97-02 49
-------
for each phase and the study. During Phase 1, only 1 of 21 pairs did not meet the 20 percent
criterion for PM10; however, only half of 10 pairs met the criterion for PM2.5. The mean bias
was about 0.5 percent for PM10 for two sites and about 28.7 percent for PM2.5 at one site. As
discussed in Sections 3.4 and 4.4, there is an apparent bias in PM2.5 between the saturation
samplers and reference dichotomous samplers. During Phase 2, 4 of 13 PM10 pairs and all 7
PM2.5 pairs did not meet the 20 percent criterion. The mean bias was about 7.9 percent for two
PM10 sites and about 27.2 percent for the one PM2.S site.
CH-97-02 50
-------
5.0 CONCLUSIONS AND RECOMMENDATIONS
Conclusions and recommendations fall into three main categories: (1) PM10 results and
their implications for monitor siting, (2) PM2.S results and their implications for monitor siting
and future studies, and (3) issues concerning the performance of the saturation samplers.
5.1 PM10 Results and Implications
The major conclusion obtained from the results and discussion in Sections 3.1.4.1 and 4.2
for PM10 may be summarized as:
• Temporal (day-to-day) variability was greater than spatial (site-to-site) variability during
both phases of the study.
• PM10 samplers in the Harlem and Washington Heights communities generally measure
lower PM10 concentrations than either of the two reference sampler sites in midtown and
downtown traffic locations.
• On average, community sites measured PM10 concentrations that were at most two-thirds
to three fourths the PM10 levels at the reference sites.
• During both phases of the study, only the two reference sampler sites recorded PM10
concentrations that indicate a likelihood of exceeding the annual PM10 standard of
50 ug/m3.
The highest PM10 levels seem to occur at those sites with the greatest traffic, which likely
explains why the two reference sites which have higher volume of bus and/or truck traffic have
higher concentration. PM10 concentrations would be expected to be high in areas with a lot of
traffic, regardless of where in the city that occurs. However, similar high PM10 levels may have
different impacts depending on the community. For example, some communities may have more
air conditioning that would reduce the impact of ambient PM10 levels; other communities may
have housing and businesses that tend to have open windows. The area around the midtown
reference sampler consists of predominantly air conditioned office buildings; however, the area
around the downtown reference sampler consists of small stores and apartments without air
conditioning. Also, in the Harlem and Washington Heights communities residents may spend
more time along heavily trafficked roads because shopping areas tend to be located along such
thoroughfare.
With limited resources, it makes sense to locate the monitors at sites where the maximum
PM10 concentrations are expected to occur. For Manhattan, this would mean keeping the current
reference sites. Within the two communities, PM10 concentrations were relatively uniform. In
Harlem, during Phase 1, the average PM10 concentration was about 36 ug/m3 for the eight
community sites with a variability among the sites of about 11 percent. The maximum
concentration site was about 13 percent above the mean while the minimum concentration site
was about 21 percent below the mean. If an additional site was to be located in the Harlem
community, it should be in the vicinity of Site 3, the Phase 1 community site with the highest
CH-9J-O2 51
-------
concentrations, located at 125th Street and Amsterdam Avenue. During Phase 2, the average
PM10 concentration was about 31.5 ug/m3 for all the community sites with a variability of about
19 percent. The maximum concentration was about 14 percent above the mean while the
minimum concentration site was about 30 percent below the mean. If an additional site was
located in the Washington Heights community, it should be in the vicinity of Site 2, located at
179th Street and Broadway, which by relative site rankings had the highest concentrations.
52 PM15 Results and Implications
Now that EPA has proposed a new PM2.5 standard," the PM2.5 results from this study
assume even greater significance. The major conclusions for PM2.5 may be summarized as
follows:
• Many sites in both communities seem likely to fail to meet the proposed annual PM2.5
standard of 15 ug/m3.11
• A few sites in both communities may also fail to meet the proposed 24-hour PM2.5
standard of 50
• Most of these sites will fail to meet the proposed 24-hour and/or annual PM2.5 standards
regardless of whether the reference sampler value (0.66) or saturation sampler value
(0.78) is used for the PM2.5/PM10 ratio is used to estimate PM2.5 from PM10.
• PM2.5 concentrations are likely to be less at sites within the two communities than at
either of the two reference sites.
A detailed quantitative estimate of the likelihood of exceeding the proposed PM2.5
.standards was beyond the scope of this project. The proposed annual PM2.S standard is based
on annual averages of eligible sites within Spatial Averaging Zones. Historically, the summer
season has PM10 and PM2.5 concentrations about 20 percent higher than the annual average.
For the period of this study, however, the weather seemed milder, with fewer stagnation periods
than normal. In general, the summer for this project would be expected to have lower paniculate
concentrations than the historical summer average.
The proposed 24-hour PM2.5 standard is based on the 98th percentile of 24-hour values
over a 3-year period. For an everyday sampling schedule, and no missing data, one would look
at the 22nd highest 24-hour value over 3 years and for every sixth day sampling, one would look
at the highest 24-hour value over 3 years. For a typical year, and no missing data, one would
look at the 8th highest 24-hour value for everyday sampling and at the second highest 24-hour
value for once every sixth day sampling to estimate if the standard is going to be exceeded. The
proposed standards give the detailed procedures for calculating the values to compare against the
standards.
If the proposed PM2.5 standards are adopted, both the City and State will need to address
the issue of how to control fine paniculate emissions. The three likely major source categories
are diesel emissions from buses and trucks, automotive emissions, and transport of secondary
CH-97-02 52
-------
aerosols. A source characterization and source apportionment study will likely need to be done
to address how much of the PM2.5 is attributable to these sources. Unlike previous studies, this
apportionment study should cover at least an entire year to detect seasonal differences in the
source characterization and apportionment and should be performed at different sites to reflect
the different mix of automobile, bus, and truck traffic that was found at various sites. The
• highest PM10 sites in the Harlem and Washington Heights communities would be good candidate
locations fo be considered for such a study. A prime objective of the study should be to
determine to what extent diesel emissions affect PM2.5 and to what extent control measures such
as converting buses to compressed natural gas are likely to reduce ambient PM2.5 levels.
>
5.3 Performance of Saturation Samplers
4
„ The saturation samplers generally gave results comparable to those of the reference
samplers for PM10. However, the samplers were biased about 28 percent high for PM2.S values
as determined by the dichotomous samplers used as PM10 reference samplers. The dichotomous
samplers are not a reference or equivalent method for PM2.S. In a previous study in Philadelphia
in 1994, the saturation samplers gave results comparable to an earlier Harvard study14 at some
of the same Philadelphia sites. The saturation samplers had determined a PM2.5/PM10 ratio of
0.76 while Harvard, using samplers of its own design, had obtained a PM2.5/PMK) ratio of
0.75.12-au
No reference or equivalent samplers for PM2.5 yet exist because EPA has only proposed,
not promulgated, a PM2.5 standard and reference method. At the current time, the first prototype
•-Federal Reference Method PM2.S samplers are just beginning to be evaluated in the field. The
SMR will soon include a set of PM2.S saturation samplers in one of the first evaluations which
is to be held at the Research Triangle Park beginning mid-December 1996. Dichotomous
samplers will also be included in the evaluation. Hopefully, this evaluation will determine if
PM2.5 saturation sampler measurements are comparable to the proposed FRM.
As discussed in Section 2.3, the saturation samplers had some operational problems such
as failing to sample for the time set. The SMR was unable to positively diagnose a cause to
account for most of the failures because of multiple component failures that affect performance.
Although many samplers were checked, repaired, and checked again, they would continue to fail
in the field and seem fine in the laboratory. The vast majority of failures in the field did not
seem to be due to operator error, flow cutoff adjustments, or batteries. Field conditions such as
vibration and electronic interference were suspected as causes, but eliminated because some
samplers would operate successfully at the same sites. The SMR is continuing to completely
overhaul the failed units. In addition, the SMR has prepared a series of stress tests to check the
failed samplers that have been repaired.
CH-97-Q2 53
-------
6.0 ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of Professor Stanley Gedzelman of the
City College of New York in providing space for the field laboratory and for supervising the
graduate and undergraduate students serving as field operators for the study. In addition, the
authors wish to thank Professor Patrick Kinney of Columbia University's School of Public Health
for sharing some of his preliminary data with us.
CH-97-02 54
-------
7.0 REFERENCES
1. AirMetrics, MiniVOL Portable Sampler Operation Manual, Springfield, OR, June 1993.
2. TRC Environmental Corporation, Quality Assurance Project Plan, Manhattan Community
Based Paniculate Study, Chapel Hill, NC, July 1996. Revised September 1996.
3. J.C. Chow, "Measurement Methods to Determine Compliance with Ambient Air Quality
Standards for Suspended Particles," Journal of the Air and Waste Management
Association, 45(5):320 (1995).
4. "Ambient Air Quality Surveillance," Code of Federal Regulations, Title 40, Part 58,
Appendix A, July 1994.
•
5. Personal communication from Prof. P.L. Kinney, Columbia University to M. Mustafa,
EPA Region D, November 1996.
6. D.N. Wittorff, A.W. Gertler, J.C. Chow, el al, "The Impact of Mobile Source Emissions
on Ambient Paniculate Loadings at a Midtown Manhattan Site," paper prepared for
publication.
7. "National Primary and Secondary Ambient Air Quality Standards," Code of Federal
Regulations, Title 40, Part 50, Section 50.6 and Appendix K, July 1994.
8. M.R. Spiegel, Schaum 's Outline of Theory and Problems of Statistics, Schaum, New
York, 1961.
9. R.G.D. Steel and J.H. Torrie, Principals and Procedures of Statistics, McGraw Hill, New
York, 1960.
10. R.C. Campbell, Statistics for Biologists, Cambridge University Press, Cambridge, 1967.
11. Federal Register, Vol. 61, No. 241, December 13, 1996, pp. 65638-65713.
12. TRC Environmental Corporation, 7994 Philadelphia PM2.5 and PMJO Saturation Study,
October 1995.
13. R.J. Tropp, S.F. Sleva, W. Ramadan, et al, "Results of the 1994 Philadelphia PM2.5 and
PM10 Saturation Study," Paper 96-MP3.03, 89th Annual Meeting of the Air and Waste
Management Association, Nashville, TN, June 23-28, 1996.
14. R.M. Burton, H.H. Suh, and P. Koutrakis, "Spatial Variation in Paniculate Concentrations
within Metropolitan Philadelphia," Environ. Sci. Technology, 30:400 (1996).
CH-97-02 55
-------
TECHNICAL REPORT ABSTRACT
REPORT TITLE Manhattan Community Based Paniculate Study
REPORT DATE January 15. 1997
CONTRACT NO. EPA 68-D3-0029
PRIME CONTRACTOR TRC Environmental Corporation
WORK ASSIGNMENT NONDELIVERY ORDER NO. (if applicable) WA3-115
PROJECT OFFICER Marv Wilkins (MD-15)
PROJECT OFFICER ADDRESS Research Triangle Park. NC 27711
PROJECT OFFICER TELEPHONE (919) 541-5229
PROGRAM OFFICE U.S. EPA Office of Air Quality Planning and Standards
NO. OF PAGES IN REPORT 100 pages
DOES THIS REPORT CONTAIN CONFIDENTIAL BUSINESS INFORMATION
YES NO X '
REPORT ABSTRACT
This report presents the results of a study to examine PM10 concentrations in two communities
in northern Manhattan over a period of approximately 80 days during July through September
1996. Sampling was conducted using saturation samplers and was divided into two phases - the
first at eight street level locations in the Harlem community and then at seven locations in the
Washington Heights community (plus one continuing site in the Harlem community). During
both phases, saturation samplers were located at sites in midtown and downtown Manhattan at
which the State operated standard dichotomous samplers. In addition, saturation samplers were
also used at a subset of sites to collect PM2.5 samples and, during Phase 2. samples for analysis
of elemental and organic carbon. Temporal variability for PM10 was greater than spatial
variability and the community sites generally had lower concentrations than either of the two
reference sites. Based on the ratio of PM2.5/PM10 measured at reference and community sites.
estimates of PM2.5 from PM10 measurements show that many sites might not meet the proposed
annual and/or 24-hour PM2.5 national ambient air quality standards.
KEY WORDS/DESCRIPTORS
paniculate matter monitoring. PM10. PM2.5. saturation sampler
-------
APPENDIX A
DAILY PM10 AND PM2.5 CONCENTRATIONS BY SITE FOR PHASE 1
CH-97-02
-------
Manhattan Community Based Particulate Study
Daily PM10 and PM2.5 Concentrations for Phase 1 Site 1
160
I 80
-------
Concentration (ug/m3)
•P*. oo ro o>
o o o o o
r~\
Manhattan Community Based Particulate Study!
Daily PM1Q Concentrations for Phase 1 Site 2 |
- m
m ^ "li m
m -
i i i i
I
• PM10 |
) 10 20 30 :• 40 50
Qamnlinn Plnv
Odl 1 ILJIII ly L/dy • L
Site 2 : 145th & Broadway |
•
-------
Manhattan Community Based Particulate Study
Daily PM10 Concentrations for Phase 1 Site 3
O>
O
IV3
O
Concentration (ug/m3
4^ 00
O O
PM10
10
20 30
Sampling Day
40
50
Site 3: 125th & Amsterdam |
-------
Manhattan Community Based Particulate Study
Daily PM10 Concentrations for Phase 1 Site 4
160
120
80
c
CD
O
O
40
10
20 30
Sampling Day
40
PM10
50
Site 4: Lennox & 148th
-------
160
^ 120
CO
E
c
i so
cd
"c
0)
o
c
o
° 40
0
Manhattan Community Based Particulate Study
1
Daily PM10 and PM2.5 Concentrations for Phase 1 Site 5 |
-
-
_
-
•M
^^»^^^
^^^*&s \
•QK /4 •§
>v ^^ ^^
HI N/ t y
- ^^ •• nHI ^^ I /X ••
-«> Pi ^Jj/ *• ^-
1 1 1^1
• PM10
• PM2.5
0 10 20 30 40 50
^^ •• v^
sampling uay Sjte 5. 135th & Lennox|
-------
Manhattan Community Based Particulate Study I
Daily PM 10 Concentrations for Phase 1 Site 6 |
160
|120
c
0)
u
o
0 40
0
C
: J?
I
- ^p ^ / «P I *1 "T|||i
) 10 20 30
Sampling Day
-
• PM10 |
40 50
Site 6: Edgecombe & 139th (Control))
-------
Concentration (ug/m3)
__ ** CD IO" CO
o o o o o
*-»
Manhattan Community Based Particulate Study
Daily PM 10 Concentrations for Phase 1 Site 7
~
-j"r* L • Jy in^ip
'
• PM10 |
-
) 10 20 30 40 50.
Sampling Day
t
[Site 7: Lexington* 99th |
-------
Manhattan Community Based Particulate Study |
Daily PM10 Concentrations for Phase 1 Site 8
160
1
120
"1 80
CD
O
c
O
O 40
PM10
10
20 30
Sampling Day
40
50
Site 8:116th & Lexington
-------
Manhattan Community Based Particulate Study
Daily PM10 and PM2.5 Concentrations for Phase 1 Site 9
160
c
0)
o
o
0 40
10
20 30
Sampling Day
O
O
40
PM10
PM2.5
50
Site 9: Madison Ave. (Ref.)|
-------
Manhattan Community Based Particulate Study
Daily PM10 Concentrations for Phase 1 Site 10
160
I 120
^)
80
CD
o
40
10
PM10
20 30
Sampling Day
40
50
Site 10: Canal St. (Ref.) I
-------
APPENDIX B
DAILY PM10 AND PM2.5 CONCENTRATIONS BY SITE FOR PHASE 2
CH-97-02
-------
100
80
CO*
1 60
c
0
1
8 40
C
O
O
20
0
C
Manhattan Community Based Particulate Study I
Daily PM10 and PM2.5 Concentrations for Phase 2 Site 1 |
•»
•
/ \ •
• H • \ 7\
• °iui P
o* ^ ^J**
ny o
iiit
-
• PM10
* PM2.5
) .10 20 30 40 50
Sampling Day
[Site 1: 181st & St. Nicholas!
-------
CO
•
s
§
u
c
o
O
100
80
60
40
20
Manhattan Community Based Particulate Study]
Daily PM10 and PM2.5 Concentrations for Phase 2 Site 2
10
20 30
Sampling Day
40
PM10
PM2.5
50
Site 2:179th & Broadway|
-------
Manhattan Community Based Particulate Study
Daily PM10 Concentrations for Phase 2 Site 3
100
80
cr?
E
"B>
3. 60
.g
*j3
2
•^
*-*
§ 40
o
c
0
O
20
0
-
—
•
~1 !•
/ / ^^
• I \ •
HI / 1 j
/ \ / I H /
/ \ • 1 J\ i
(B I^B / I p^ i i
i^ "mH / H^H^ \ i
/ / ^^^81 \ /
• rfV 1. /
VA
i i i i
• PM10 |
0 10 20 30 , 40 50
Sampling Day
[Site 3: 168th & St. Nicholas!
-------
CO
c
o
•*=
2
4->
0>
o
c
o
O
100
80
60
40
20
Manhattan Community Based Particulate Study |
Daily PM10 Concentrations for Phase 2 Site 4
PM10
J
10
20 30
Sampling Day
40
50
Site 4:162nd & Edgecombe (Control)!
-------
100
80
S. 60
c
o
g
§ 40
u
c
o
O
20
Manhattan Gommunity Based Particulate Study
Daily PM10 Concentrations for Phase 2 Site 5
V>
10
20 30
Sampling Day
40
50
PM10
]
Site 5: Dyckman & Sherman |
-------
Manhattan Community Based Particulate Study]
Daily PM10 Concentrations for Phase 2 Site 6
100
80
1 60
c
o
Concentra
*k
o
ro
o
PM10
10
20 30
Sampling Day
40
50
Site 6: 207th & 9th Ave.
-------
100
80
CO*
Concentration (ug/
£t O)
0 0
20
0
C
Manhattan Community Based Paniculate Study
Daily PM10 Concentrations for Phase 2 Site 7
i V " ]l
• l\f m
• PM10 |
) 10 20 30 40 50
Sampling Day
Site 7: 214th & Broadway |
-------
Manhattan Community Based Particulate Study]
Daily PM10 Concentrations for Phase 2 Site 8
100
80
CO
60
40
u
o
O
20
10
PM10
20 30
Sampling Day
40
50
[Site 8:125th & Amsterdam (Phase 1 Site 3)|
-------
Manhattan Community Based Particulate Study]
Daily PM10 and PM2.5 Concentrations for Phase 2 Site 9
100
80
CO
1 60
c
o
u
o
O
40
20
O
O
10
20 30
Sampling Day
40
PM10
PM2.5
50
[Site 9: Madison Ave. (Ref.)|
-------
100
80
CO
I
60
-------
APPENDIX C
COMPARISON OF PM10 AND PM2.5 CONCENTRATIONS MEASURED BY
REFERENCE AND SATURATION SAMPLERS
CH-97-02
-------
Manhattan Community Based Participate Study
Phase 1 - PM10 Concentrations at Reference Site 9 (Madison Ave.)
160
140
co* 120
Concentration (ugy
£fc O) 00 O
O O O O
20
0
C
•
O
•
o •
- - . 9° ?
• • ^ ^
™ • n £ii
D
i i i i
-
• PM10(Ref.)
^PM10(Sat.)
) 10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Particulate Study
Phase 1 - PM10 at Reference Site 9 (Madison Ave.)
160
I
u
120
100
80
Q.
I 60
CO
20
0
20 40 60 80 100 120 140 160
Reference Sampler PM10 Cone. (ug/m3)
• PM10 Data
PMIORegr.
-1:1 Line
[S]« 0.9581R] + 1.043
r2 = 0.92 (less outlier)
-------
Concentration (ug/m3)
Manhattan Community Based Participate Study
Phase 1 - PM10 Concentrations at Reference Site 10 (Canal St.)
160
140
120
100
80
60
40
20
o I
C
: • .« *
W £ wO
• \s B <^™" «
• 0 I «
a
8
II 1 1
• PM10(Ref.)
<|>PM10(Sat.)
) 10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Particulate Study
Phase 1 - PM10 at Reference Site 10 (Canal St.)
CO
100
80
o
o
O
o 60
_
Q.
as
CO
c
40
2 20
CO
20 40 60 80
Reference Sampler PM10 Cone. (ug/m3)
100
• PM10 Data
PMIORegr.
-1:1 Line
1.040*[RJ + 0.456
r2 = 0.761
-------
Manhattan Community Based Participate Study I
Phase 1 - PM2.5 Concentrations at Reference Site 9 (Madison Ave.) |
100
80
CO
^>
3, 60
c
0
Concentrat
^
o
20
0
C
o
00"
— mt
/^ •
^ • ^
• ^
• • 0
• . • *' .
• Q B "
•
i i i • i
• PM2.5 (Ref.)
^ PM2.5 (Sat.)
) 10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Participate Study]
Phase 1 - PM2.5 at Reference Site 9 (Madison Ave.)
100
CO
o> 80
60
o
o
O
in
c\i
Q.
J5
a.
cd
CO
c
.g
1 20
4->
CO
40
PM2.5 Data
PM2.5 Regr.
1:1 Line
+ 4.842
r2 . Q.921
20 40 60 80
Reference Sampler PM2.5 Cone. (ug/m3)
100
-------
Manhattan Community Based Particujate Study
Phase 1 - PM2.5 Concentrations at Reference Site 10 (Canal St.)
Concentration (ug/m3)
I\D 4^ O) CO O
O O O O O O .
«-» I
• •
• • * ** •
•
I 1 1 I
• PM2.5 (Ref.) |
) 10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Particulate Study
Phase 1 - PM2.5 from Reference Samplers
CO
75
c
CJ.
O
100
80
CD
55 6°
CO
o
o
o
in
cvi
40
20
0
• PM2.5 Data
PM2.5 Regr.
— 1:1 Line
0 20 40 60 80 1Q
PM2.5 Cone. (ug/m3) at Site 9 (Madison Ave.)
[10] = 0.879604*[9] + 5.683366
r2 = 0.814628
-------
Manhattan Community Based Particulate Study
Phase 2 - PM10 Concentrations at Reference Site 9 (Madison Ave.)
co
100
80
60
• 40
O
c
O
O
20
O
PMlO(Ref.)
PM10(Sat.)
10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Particulate Study
Phase 2 - PM10 at Reference Site 9 (Madison Ave.)
CO
100
80
o
c
o
O
o 60
2
Q_
|- 40
05
CO
c
o
'•§ 20
13
03
CO
0
20 40 60 80 100
Reference Sampler PM10 Cone. (ug/m3)
• PM10 Data
PMIORegr.
— 1:1 Line
= 0.978 (less outlier)
-------
Manhattan Community Based Particulate Study
-
Phase 2 - PM1 0 Concentrations at Reference Site 1 0 (Canal St.)
100
80
CO*
E
2- 60
c
0
'1
^BM A *%
c 40
0)
o
c
o
O
. 20
0
•«•
<^
i •
m A ®
'
o • o *
A $> O
•
^ ^ •
H '
g
i i i i
• PM10(Ref.)
^PM10(Sat.)
0 10 20 30 40 50
Sampling Day
-------
Manhattan Community Based Particulate Study]
Phase 2 - PM10 at Reference Site 10 (Canal St.)
100
CO
80
d
o
O
o 60
40
to
CO
g 20
13
CO
20
40 60 80
Reference Sampler PM10 Cone. (ug/m3)
100
• PM10 Data
PMIORegr.
-1:1 Line
IS] = 0.725*(R] + 17.48
r2 = 0.688
-------
Manhattan Community Based Particulate Study
Phase 2 - PM2.5 Concentrations at Reference Site 9 (Madison Ave.)
100
o
O 20
0 L
PM2.5 (Ref.)
PM2.5 (Sat.)
10 20 30 40
Sampling Day
50
-------
Manhattan Community Based Particulate Study
Phase 2 - PM2.5 at Reference Site 9 (Madison Ave.)
100
OT*
^)
2- 80
•
o
60
40
o
O
in
cvi
OL
t_
_0>
Q.
(d
CO
c
I 20
3
OS
CO
• PM2.5 Data
PM2.5 Regr.
-1:1 Line
[SJ=1.040*[R]-i-9.007|
r2 = 0.988 (less outlier)!
20 40 60 80
Reference Sampler PM2.5 Cone. (ug/m3)
100
-------
Manhattan Community Based Particulate Study
Phase 2 - PM2.5 Concentrations at Reference Site 10 (Canal St.)
100
80
CO
s
+-*
0)
o
c
o
O
60
40
20
0
0
10
20 30
Sampling Day
40
PM2.5 (Ref.)
50
-------
Manhattan Community Based Particulate Study]
Phase 2 - PM2.5 from Reference Samplers
100
w 80
73
5
o,
o
2 60
CO
CO
d
in
cvi
40
20
• PM10 Data
PMIORegr.
- 1:1 Une
(10] - 0.978183*[9] + 1.873696
r2 - 0.420183
20 40 60 80
PM2.5 Cone. (ug/m3) at Site 9 (Madison Ave.)
100
------- |