EPA/600/R/12/621 February 2014 | www.epa.gov/ord
United States
Environmental Protection
Agency
Cicero Rail Yard Study (CIRYS)
Final Report
Office of Research and Development
-------
EPA/600/R-12/621
February 2014
Cicero Rail Yard Study (CIRYS)
Final Report
Michael Rizzo, Jesse McGrath, Chad McEvoy, Marta Fuoco
US EPA Region 5, Chicago, IL
Gayle Hagler, Eben Thoma
US EPA ORD, Research Triangle Park, NC
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
Office of Research and Development
US Environmental Protection Agency
Cincinnati, OH 45268
-------
Table of Contents
Executive Summary 10
1. Introduction 12
1.1. Background 12
1.2. Purpose and Scope of Study 14
2. Methods 14
2.1. Site Selection Process and Description 14
2.2. Mobile and Stationary Sampling Approach and Schedule 16
2.3. Mobile Monitoring Methods 17
2.3.1. Mobile Monitoring Instrumentation 17
2.3.2. Driving Route 19
2.3.3. Mobile Data Processing 21
2.3.4. Quality Assurance 25
2.4. Stationary Sampling Methods 26
2.4.1. Instrumentation 26
2.4.2. Stationary Monitoring Location 27
2.4.3. Data Processing 28
2.4.4. Quality Assurance 31
2.5. Ancillary data 31
3. Data Analysis 32
3.1. Mobile Sampling 32
3.1.1. Mobile Data Overview 32
3.1.1.1. Quality assurance review and data completeness 34
3.1.1.2. Sampling sessions in the context of meteorology 37
3.1.1.3. Sampling sessions in the context of rail yard activity 40
3.1.1.4. Mobile / Stationary data comparison 43
3.1.2. Assessment of local air quality impact through mobile monitoring 46
3.1.2.1. Downwind and upwind comparison 47
3.1.2.2. Wind speed effect 52
3.1.2.3. Impact as function of distance 53
3.2. Stationary Sampling 55
3.2.1. Stationary Data Overview 55
3.2.1.1. Quality assurance review and data completeness 55
3.2.1.2. Data collection in context of local meteorology 58
3.2.2. Assessment of rail yard impact 60
3.2.2.1. Impacts of meteorology and time of day on air quality measurements 60
3.2.2.2. Nonparametric trajectory analysis of stationary monitoring data 66
4. Summary and Conclusions 71
5. Acknowledgements 72
6. References 73
-------
List of Appendices
Appendix A. Quality Assurance Project Plans
Appendix B. Mobile Monitoring Quality Check Results
Appendix C. Mobile and Stationary Side-by-Side Sampling
Appendix D. Mobile Monitoring Wind Roses and Driving Maps
-------
List of Figures
Figure 1-1. Ultrafine particles (UFP) as a function of downwind distance from
the roadway, with concentrations normalized by the location nearest to the road.
Originally published in Hagler et al., (2009). 12
Figure 2-1. Annual number of container lifts per rail yard facility in the Chicago area.
Data source: Chicago Metropolitan Agency for Planning (2011). The rail yard focused
on for this research, Cicero, is shaded in blue. 16
Figure 2-2. Mobile sampling vehicle driving route surrounding the rail yard of interest. 20
Figure 2-3. Mobile data processing steps. 21
Figure 2-4. Example time series of carbon monoxide data, showing the original raw data
(gray) and the data time series after detected significant short-term spikes were removed,
which were presumably due to local gasoline vehicle exhaust impacting the measurements. 23
Figure 2-5. Example of distance estimated from the pathway driven in neighborhood
NT1 relative to the rail yard edge. The dashed line demonstrates the shortest distance from
an example location along the driving route to the estimated rail yard boundary line.
The shortest distance from each location in neighborhood NT1 to the rail yard boundary
is calculated and can be compared to the color bar at the top of the figure. 24
Figure 2-6. Example of a concentration versus distance from rail yard boundary figure for
CO (left) and black carbon (right). The circle markers indicate the mean value measured
and are centered at the midpoint of the distance range; the dashed line indicates the
standard error of the measurement. 25
Figure 2-7. Sampling location, where the stationary monitoring site was located for
approximately a year's time and the mobile vehicle would park for 1-2 hour
intercomparison periods. 27
Figure 2-8. Example wind data showing how a typical backward trajectory is constructed
for NTA. 29
Figure 2-9. Example showing all of the backward trajectories for all 5 minute averaged
data collected at the Cicero stationary monitor 30
Figure 3-1. Driving speed recorded by the sampling vehicle's GPS 33
Figure 3-2. Picture of the instrumentation in the vehicle (left) and sampling vehicle
in action (right) at the Cicero Rail Yard. 33
Figure 3-3. Example particulate time series during the addition of a zero filter and after
the filter is removed. 35
-------
Figure 3-4. Comparison of wind measurements collected at Midway airport (left) with wind
data collected on top of a BNSF building (right). 36
Figure 3-5. Comparison of wind measurements collected at the stationary monitoring
location (left) with wind data collected on top of a BNSF building (right). 37
Figure 3-6. Morning session wind trends - arrow orientation indicates wind direction
(e.g., pointing towards N means wind from the S) and extent indicates mean wind
speed. Sessions shown are #5, 8, 10, 14, 16, 20, 21, and 22. 39
Figure 3-7. Mid-day session wind trends - arrow orientation indicates wind direction
(e.g., pointing towards N means wind from the S) and extent indicates mean wind speed.
Sessions shown are 1, 4, 7, 9, 13, 15, and 19. 39
Figure 3-8. Evening session wind trends - arrow orientation indicates wind direction
(e.g., pointing towards N means wind from the S) and extent indicates mean wind speed.
Sessions shown are #2, 3, 6, 11, 12, 17, 18, and 23. 40
Figure 3-9. Daily total crane container lifts (a) and truck counts at the gate (b)
during the mobile sampling period of October 27, 2010 through November, 22, 2010. 41
Figure 3-10. Average gate counts (left) and lift counts (right) by day of week and
hour of day during the study period of October 27, 2010 through November, 22, 2010. 42
Figure 3-11. Comparison of mean diurnal gate activity during the study
(heavier dashed black line) with monthly average diurnal trends from
October 2010 to July 2011 (10 months) shown in thin colored lines. 42
Figure 3-12. Comparison of mean diurnal lift activity during the study
(heavier dashed black line) with monthly average diurnal trends from
October 2010 to July 2011 (10 months) shown in thin colored lines. 43
Figure 3-13. Wind rose during a period of mobile and stationary site side-by-side sampling. 44
Figure 3-14. Parallel time series of concentrations for the stationary monitoring site
(green) reporting raw data at 5- minute intervals and the real-time data collected
onboard the mobile monitoring vehicle (blue). Note: Lower limit of detection for the
stationary CO analyzer is 300 ppb. 45
Figure 3-15. Black carbon time series for the mobile car (real-time in dark blue, 5-minute
average in light blue) and 5-minute stationary data (green). 46
Figure 3-16. Statistically significant excess concentrations above the background
for neighborhood transects NT1-NT4, calculated for areas up to 300 m from the rail yard. 49
-------
Figure 3-17. Fraction of neighborhood areas (NT1-NT4), in 50 m increments up to
300 m from the rail yard, with significant increase in pollutant levels above the
background during early morning sessions (~4-7 AM). 50
Figure 3-18. Fraction of neighborhood areas (NT1-NT4), in 50 m increments up to
300 m from the rail yard, with significant increase in pollutant levels above the
background during mid-afternoon time periods (~10 AM-1 PM). 51
Figure 3-19. Fraction of neighborhood areas (NT1-NT4), in 50 m increments up to
300 m from the rail yard, with significant increase in pollutant levels above the
background during evening time periods (7-10 PM). 52
Figure 3-20. Mean downwind BC concentrations in neighborhood areas up to 300 m
from the rail yard, as a function of local wind speed. 53
Figure 3-21. Normalized downwind excess BC during early morning sampling sessions
in neighborhood areas up to 300 m from the rail yard, as a function of distance from
the rail yard. 54
Figure 3-22. Normalized downwind excess BC during evening sampling sessions in
neighborhood areas up to 300 m from the rail yard, as a function of distance from
the rail yard. 54
Figure 3-23. Time series and histogram summarizing stationary monitoring data
collected during the time period isolated for analysis (November 2010 - May 2011). 57
Figure 3-24. Wind trend comparison for the Cicero monitoring site compared with
the Midway airport meteorological station. 59
Figure 3-25. Cicero stationary monitoring site wind trends during the day (left figure)
and night (right figure). 60
Figure 3-26. Concentration roses for sulfur dioxide (a), nitrogen dioxide (b),
black carbon (c), and nitrogen dioxide (d). 61
Figure 3-27. Diurnal concentration roses for sulfur dioxide (a), nitrogen dioxide (b),
black carbon (c), and nitrogen dioxide (d). 62
Figure 3-28. Normalized diurnal time series of measured concentrations at the
stationary monitoring site, for all data collected during November, 2010 - May, 2011. 63
Figure 3-29. Normalized diurnal time series of measured concentrations at the
stationary monitoring site, for weekday (left) and weekend periods (right) collected
during November, 2010-May, 2011. 64
Figure 3-30. Normalized diurnal time series of measured concentrations at the
-------
stationary monitoring site, for periods of wind from the North (left) and wind from
the South (right) collected during November, 2010 - May, 2011. 65
Figure 3-31. Black carbon expected concentration field from NTA in ng/m3 67
Figure 3-32. Sulfur dioxide expected concentration field from NTA. 68
Figure 3-33. Nitric oxide expected concentration field from NTA. 69
-------
List of Tables
Table 2-1. Site selection criteria
Table 2-2. Measurement approaches
Table 2-3. Mobile monitoring instrumentation
Table 2-4 Stationary monitoring instrumentation
Table 2-5. Stationary monitoring site quality assurance procedures
Table 3-1. Mobile Sampling Sessions
Table 3-2. Example QC metrics for the air monitoring instruments onboard the
sampling vehicle
Table 3-3. Data completeness
Table 3-4. Wind characteristics during mobile sampling sessions
Table 3-5. Stationary monitoring data completeness by month
T ,, _ _ Summary statistics for 5-minute pollutant data
Table 3-6. y ^
(NOx and SO2 in ppb, BC in ng m~3)
Table 3-7. Pearson Correlation Coefficients (R) for Pollutant Pairs
15
17
19
26
31
32
35
35
38
56
58
70
-------
List of Acronyms
Acronym
APS
BAM
BC
BGD
CIRYS
CO
DQO
EC
EEPS
FEM
FRM
GPS
HEPA
NAAQS
NAMS
NCDC
NOAA
NDIR
NO
NO2
NT
NTA
OC
ONA
ORD
PM2.5
PM10
QA
QAPP
RARE
SLAMS
UFP
USEPA
Description
Aerodynamic Particle Sizer
Beta-Attenuation Mass monitor
Black carbon
Background
Cicero Rail Yard Study
Carbon monoxide
Data Quality Objective
Elemental carbon
Engine Exhaust Particle Sizer
Federal Equivalent Method
Federal Reference Method
Global Positioning System
High-Efficiency Particulate Air (type of filter)
National Ambient Air Quality Standards
National Air Monitoring Sites
National Climatic Data Center
National Oceanic and Atmospheric Administration
Non-dispersive Infrared (type of detector)
Nitrogen oxide
Nitrogen dioxide
Neighborhood Transect
Non-parametric Trajectory Analysis
Organic carbon
Optimized Noise-reduction Algorithm
Office of Research and Development
Particulate matter smaller than 2.5 microns
Particulate matter smaller than 10 microns
Quality assurance
Quality assurance project plan
Regional Applied Research Effort
State and Local Air Monitoring Stations
Ultrafine particle
United States Environmental Protection Agency
-------
The Cicero Rail Yard Study (CIRYS) was initiated as a second phase of an EPA Region 5 Regional Applied
Research Effort (RARE), a program which facilitates the collaboration between a particular Region and
EPA's Office of Research and Development (ORD) in an area of research where the Region needs
scientific information to support implementation of a specific program or strategy. EPA Region 5
identified local scale air pollution impacts due to rail yard emissions as a poorly understood issue and
the region's concern was solidified by a regulatory monitoring station measuring elevated readings for
PM2.5 at a location near a rail yard in Dearborn, Ml. To better understand the local-scale air pollution
impact of rail yard activities, the RARE effort covered two independent research studies - Phase I, the
emissions inventory, modeling, and field measurements conducted at a rail yard in Dearborn, Michigan,
and Phase II, a two-pronged field measurement campaign conducted near a rail yard in Cicero, Illinois.
This report covers this Phase II effort in Cicero, IL
Field measurements of local scale air pollution are an evolving field of research and require advances in
air quality measurement. Previous near-road studies documented that air pollution concentrations
within close proximity to a highway can vary significantly in both time and space, with local meteorology
and the local terrain significantly affecting near-field concentrations of traffic-related pollutants (e.g.,
Baldauf et al. 2008a, Karner et al. 2010). The rail yard environment was considered likely to be even
more complex in nature than a highway, given emissions from multiple sources that vary in time and
location within the yard. Therefore, ideal field measurements would cover a wide spatial area as well as
a long time horizon. To meet this goal, a combined strategy of short-term mobile and longer-term
stationary monitoring was identified. A variety of rail yards in the Chicago area were surveyed to
determine which rail yard environment would be compatible with this field measurement strategy as
well as meeting several other study goals, including avoidance of potential confounding sources as well
as location in an urban environment. The Cicero rail yard was selected for study, based upon these
objectives.
The Cicero rail yard is an intermodal rail yard, with emissions including both truck and locomotive
operations. Measuring rail yard activity by freight container lifts, the Cicero rail yard ranked 7th in the
Chicago area in 2011 (out of 19 total) with 370,000 lifts. In addition to freight trains passing through the
rail yard, a commuter train line also passes along the northern border of the Cicero rail yard. The
surrounding environment is primarily residential with two-story single-family homes densely located in a
grid fashion surrounding the yard. With prevailing winds from the southwest, a stationary monitoring
site was located to the northeast of the rail yard at approximately 50 meters from the train tracks
running along the northern border of the yard and provided continuous 5-minute measurements of
sulfur dioxide, oxides of nitrogen, carbon monoxide, black carbon, fine particulate matter, and
meteorology over nearly a year timeframe. The mobile monitoring driving route was designed to
measure air pollutant concentrations along low-traffic residential roads north of the rail yard, as well as
capturing urban background air pollution levels from areas upwind of the yard. Mobile monitoring was
conducted over a one month period and utilized advanced real-time measurement instrumentation for
carbon monoxide, black carbon, particle size distribution (ultrafine to coarse range), and location.
10
-------
Key findings from the CIRYS study include the following:
• Stationary monitoring results
o Analysis of an approximate 6 month period of continuous data collection (November
2010 to May 2011) determined elevated sulfur dioxide, black carbon, and oxides of
nitrogen during conditions of wind from the southern sector.
o Diurnal analyses of stationary monitoring results, isolated in weekday/weekend
timeframes and North / South wind timeframes, indicate higher overall pollution levels
on weekdays and with winds from the south.
o Nonparametric trajectory analysis (NTA), an inverse modeling approach utilizing high
time resolution monitoring and wind data, estimated multiple southern source areas
contributing to the elevated concentrations under southerly winds, including the rail
yard area, an airport located due south, and a nearby power plant.
• Mobile measurement results
o Evaluation of mobile air monitoring sessions - with data sets including carbon
monoxide, ultrafine particles, fine and coarse particle counts, and black carbon -
indicate that residential areas north of the rail yard have elevated black carbon
concentrations relative to a residential area south of the yard (estimated 30-104%
increase over urban background) during early morning and evening time periods with
southerly winds. Other pollutant measurements on the mobile platform either did not
show statistically significant increases relative to the background or were less
consistent.
o Black carbon in the northern neighborhoods was elevated under both northerly and
southerly winds during the mid-day period relative to a residential area south of the
yard, indicating that other local sources had higher emissions activity during this time.
o During the early morning and evening sessions with southerly wind, excess black carbon
concentrations in northern neighborhoods are shown to decrease with increasing wind
speed. In addition, elevated black carbon concentrations in these downwind
neighborhoods do not appear to have a consistent trend associated with increasing
downwind distance from the rail yard boundary, which is likely related to the complexity
of rail yard and surrounding environment.
These results support the notion that local concentrated areas of higher diesel emissions activity
adversely impact local-scale air quality and mitigation efforts may reduce local exposure to air
pollution. It should be noted that uncertainty remains regarding source attribution, both within the
rail yard and considering potential traffic on boundary roads, which may require modeling or
controlled field experiments for further characterization.
11
-------
1. Introduction
1.1.Background
Near-source air quality is currently a priority research area for the US EPA, with recent research
studies focusing on transit sources such as major roads and rail yards. "Near-source" generally
constitutes areas within several hundred meters of a major source, where excess local air pollution
may be present in addition to regional background air pollution levels. The vast majority of past
near-source research has focused on the near-road environment, with very limited information to
date on rail-related environments.
Past near-road research has determined that certain air pollutants - particularly those produced
directly from the tail pipe - can have significantly higher levels in close proximity to roadways and
generally attenuate exponentially with distance from a road (Karner et al. 2010). For example,
ultrafine particles (UFPs, particles smaller than 100 nanometers in diameter) have been observed in
multiple locations to be significantly higher in close proximity to roadways and to decrease quickly
with distance (Figure 1-1). For air pollutants that are composed of constituents formed post-
emission through chemical reactions in the atmosphere and with a higher regional background, such
as fine particulate matter, the near-source effect is usually smaller but may still be elevated relative
to background concentrations. A review of health studies related to near-road exposure has
determined a significant association between residing near roads and the exacerbation of asthma;
other health outcomes (e.g., cardiovascular mortality, onset of childhood asthma) also have
suggestive associations but more research is required (HEI Panel on the Health Effects of Traffic-
Related Air Pollution 2010). Important factors that affect near-road air pollution include
meteorology, road design and surrounding structures, traffic volume, fleet mix, and driving mode
(e.g., Baldauf et al. 2008a, Hu et al. 2009).
too
i This study
Beckermsn et al (2007) Site t
Beckerman el al (2007) Site 2
Phoplaetal |20Q7)
Zhu et al (2002ai
Zhuelal(2002t>|
mavra el al (2000)
000
50 100 150 200 250
distance from roadway (m)
Figure 1-1. Ultrafine particles (UFP) as a function of downwind distance from the roadway, with
concentrations normalized by the location nearest to the road. Originally published in Hagler et al., (2009).
Limited information is currently available on near-rail air pollution. The rail environment may be
considered in two parts -the network of tracks connecting destinations and the "nodes" of the
12
-------
network (rail yards) where freight is organized and may be moved from one transportation system
to another. Within the category of rail yards, a defining characteristic is whether they exclusively
move freight containers from one train to another (classification) or whether the rail yard also
serves as a location where freight may move to other modes of transportation such as trucks
(intermodal). Emissions within a rail yard - particularly intermodal rail yards - are considerably
more complex than highway environments. While the rail yard is a fixed area, emissions within an
intermodal yard are of multiple types, including locomotives passing through the yard, switcher
locomotives that route containers within the yard, trucks coming to and from the yard, hostler
trucks moving containers within the yard, cranes moving containers, and other distributed emissions
associated with servicing the rail yard equipment (Federal Highway Association 2010). These
emissions can be considered as multiple point and line sources, which may shift both in location and
emissions strength with time.
Two recent field studies have been conducted to estimate the local-scale air pollution impact
attributed to emissions within a rail yard. One recent study in Detroit - the Midwest Rail Study -
evaluated emissions and local-scale PM2.5 trends related to the CSX Rougemere rail yard in
Dearborn, Ml (Turner 2009). An emissions inventory estimated significant improvements in rail yard
PM2.s emissions over 2007-2008 related to replacing yard switcher locomotives with Gen Set
locomotives as well as through the introduction of low sulfur fuel. Turner (2009) also reported that
a dispersion modeling exercise estimated that the Rougemere rail yard contributed 0.2 u.g m~3 PM2.s
on an annual average basis at an air monitoring station located within 150 m East of the rail yard
boundary. The impacts were primarily attributed to locomotive emissions in the yard (switcher,
arriving/departing, and through locomotives). Upwind and downwind point monitoring of
carbonaceous particulate matter - elemental carbon (EC), black carbon (BC), and organic carbon
(OC) - was also conducted for several months in 2008. The field data appeared to be significantly
affected by other nearby sources which confounded the upwind/downwind local air quality analysis.
Another major field study took place over the timeframe of 2005-2008 to document local air
pollution impacts related to emissions from the Union Pacific Rail Road J.R. Davis rail yard that is
located in Roseville, CA. The Davis rail yard is unique in having primarily locomotive emissions,
lacking the truck traffic associated with intermodal rail yards. In addition, the rail yard had
somewhat steady and predictable winds that cross over the yard during summertime, supporting
the application of time-integrated upwind/downwind measurements (Cahill et al. 2011). During
nighttime summer conditions in 2005, it was reported that NO, NOx, BC, and PM2.s exhibited
downwind concentrations that were enhanced by a factor of 21.9 (net difference: 77 ppb), 7.1 (net
difference: 103 ppb), 2.4 (net difference: 0.7 u.g m"3), and 1.5 (net difference: 4.7 u.g m"3) in
comparison to upwind measurements, respectively (Cahill et al. 2011). Over the course of the
following three years, the pollution levels at the downwind site tapered down substantially due to
emissions improvements at the Davis rail yard. By 2008, downwind-upwind concentration
differences were reduced for NO, NOx, BC, and PM2.5 by over 50% (Campbell 2009).
While the field results from Campbell (2009) and Cahill et al. (2011) support the notion that rail
yards may significantly increase local-scale air pollution, it is uncertain to what degree these findings
13
-------
translate to rail yards with different emissions composition (e.g., intermodal yards with fewer trains,
more trucks) that are located in other areas of the country with different local meteorology trends.
As Turner (2009) experienced, certain rail yards are located in close proximity to other major
sources - these multi-source environments are of significant interest in terms of local air pollution
exposure but pose a significant challenge to isolate and assess individual source contributions to
local excess air pollution.
The objective of this study was to evaluate the impact of an active rail yard on local air quality,
within an urban area in EPA Region 5. This report describes the field measurements conducted
through a month-long mobile monitoring campaign (October-November, 2010) as well as longer-
term stationary monitoring (October 2010- October 2011) adjacent to a mid-sized rail yard in
Chicago -the BNSF Cicero Rail Yard. Additional ancillary data utilized in the data analysis include a
time series of truck counts at the BNSF gate ("gate count") and a count of freight containers lifted by
the cranes within the yard ("lift count"), as recorded by BNSF employees at the Cicero yard.
Specific scientific goals of the study included:
• Measure the spatial extent of elevated local air pollution compared to the background,
downwind of a major rail yard in Region 5.
• Measure the spatial and temporal variability of near-rail yard air pollution, under different
meteorological conditions and source emission characteristics.
• Spatially attribute source area contributions to local excess air pollution.
EPA Region 5 and EPA Office of Research and Development staff came to consensus on a number of
siting criteria to support the project objectives outlined in Section 1.2. These criteria and the
relative ranking applied to Chicago-area rail yards are provided in Table 2-1. The identified criteria
were related to several goals. One goal was to employ a mobile air monitoring approach to measure
spatial gradients of air pollution, an approach that was determined to be lower cost and more
flexible in comparison to implementing multiple stationary monitoring sites. Part of the site
objectives therefore included assessing the surrounding roadway network to determine if a mobile
sampling vehicle could travel along relatively low traffic roads in close proximity to a rail yard. In
addition, with Region 5 staff providing in-kind support to the study through implementing a
stationary monitoring site, initial site selection was focused on the Chicago area where Region 5
headquarters are located.
14
-------
Table 2-1. Site selection criteria
Criteria Rank
(H:high,
M:mid,
Activity level of rail yard H
Existence of historical monitoring data at the site M
Ease of setting up a fixed sampling site and monitoring meteorology and air H
quality for several months.*
Few other nearby sources H
Capability to drive in close proximity to rail yard on multiple sides, H
particularly along axis of prevailing wind
Access to low traffic roads surrounding rail yard, to avoid biases from single H
vehicle exhaust
Characteristics of surrounding environment (residential, commercial, etc.) M
*To support collaborative monitoring effort by Region 5 staff.
A number of monitoring sites were considered throughout the Chicago-area, including the following
rail yards - Corwith, Proviso, Cicero, 59th Street, The Belt, Ashland, and Elwood. The BNSF Cicero
Rail Yard met the majority of the selection criteria in Table 3-1 when compared to the other
candidates and was selected as the optimal site. Cicero is an intermodal rail yard, with both
locomotive and truck traffic inside the rail yard boundary. A common metric for intermodal rail yard
activity is the number of shipping container lifts per day - Cicero has approximately 1000-1200 lifts
per day. To put the Cicero yard activity into perspective, Figure 2-1 shows the number of container
lifts on an annual basis for a number of yards located in the Chicago area - in 2010, Cicero was
reported to have approximately 370,000 lifts, with other rail yards in the Chicago area ranging from
below 100,000 to above 800,000 lifts (Chicago Metropolitan Agency for Planning, 2011). In addition
to emissions by diesel-powered cranes, other on-site emissions include 5-9 hostler trucks, 500
heavy-duty trucks traveling in/out of the yard, 8 daily intermodal trains, approximately 140 through
trains (~120 passenger trains, 20 mixed freight trains), and 4-5 switcher locomotives. During the
time period of this study, intermodal equipment (cranes, hostlers) used within the yard were
operating on ultra low sulfur diesel fuel (maximum sulfur content of 15 ppm). Locomotives fueled in
the area during the time of the study were using low sulfur diesel fuel (in the range of 125-400 ppm
sulfur) (Michael Stanfill, BNSF - personal communication). The surrounding environment is
primarily residential with two-story single-family homes densely located in a grid fashion
surrounding the yard.
15
-------
Blue Island
Schiller Park
Bensenville
Gateway
Landers
Calumet
63rd Street
47th
59th Street
Bedford Park
Global III
Yard Center
Canal Street
Global II
Global I
Logistics Park
Cicero
Willow Springs
Corwith
200,000 400,000 600,000
Annual Container Lifts
800,000
Figure 2-1. Annual number of container lifts per rail yard facility in the Chicago area. Data source: Chicago
Metropolitan Agency for Planning (2011). The rail yard focused on for this research, Cicero, is shaded in blue.
2.2. Mobile and Stationary Sampling Approach and Schedule
With a goal of characterizing spatial gradients of air pollutant concentrations in neighborhoods
located near the yard as well as observing long-term temporal variation, a combined mobile and
stationary monitoring approach was implemented. EPA ORD staff led the mobile monitoring
measurements and EPA Region 5 staff led the stationary monitoring measurements. These two
approaches were complementary in nature. Mobile monitoring was conducted to drive a network
of roadways surrounding the rail yard and provide data on the spatial variability of air pollutant
concentrations. Meanwhile, the stationary monitoring was conducted at a single location over
approximately one year, allowing longer-term temporal trends in near-rail yard concentrations to be
understood. These two approaches required different air pollution measurement techniques. To
collect high spatial resolution air pollution data while the mobile monitoring vehicle was in motion,
air pollution measurement techniques were employed that were capable of measuring at a very fast
rate (<10 seconds) while maintaining enough sensitivity to resolve ambient concentration levels.
The instruments employed in the mobile monitoring vehicle were therefore on the leading edge of
technology and are not commonly found in typical air monitoring networks. In contrast, a nominal
5-minute sampling time requirement was set for the stationary monitoring site to resolve air
pollutant concentrations with changing wind direction. Therefore, instrumentation that is
commonly used for regulatory air monitoring purposes and capable of data output on the order of
16
-------
minutes was applied. A general summary of these two sampling approaches and instrumentation
are provided in Table 2-2.
Table 2-2. Measurement approaches
Mobile Monitoring Vehicle Stationary Monitoring Site
Sampling times
Time span October-November, 2010 October, 2010-October, 2011
Measurement rate 1-10 seconds, driving sessions of 5 minutes, continuous data
approximately 3 hours
Measurement techniques"
Fine particulate Aerodynamic sizing, light scattering Beta-attenuation through particle-laden
matter detection, mass-estimation from size- filter, with an inlet cut at 2.5 microns
resolved number counts (FEM)
Ultrafine particles Electrical mobility sizing, detection by N/A
electrometer
Black carbon Light absorption (880 nm) through Light absorption (880 nm) through
particle-laden filter particle-laden filter
Carbon monoxide Quantum cascade laser Nondispersive infrared detector (FRM)
Sulfur dioxide Quantum cascade laser Pulsed fluorescence (FEM)
Oxides of nitrogen N/A Chemiluminescence (FRM)
aFEM means Federal Equivalent Method and FRM means Federal Reference Method.
2.3. Mobile Monitoring Methods
The mobile monitoring data set covers the time period of October 27, 2010 to November 21, 2010
and included 23 deployments in total. Air monitoring measurements were conducted using an
electric vehicle outfitted with rapid-response air sampling instrumentation. The vehicle is powered
by lithium ion batteries, which provide enough power for approximately 100 miles of driving.
However, the actual driving time depends on the true operating conditions of the vehicle, with
higher speed driving more quickly draining the power supply. Under the driving conditions in CIRYS,
the vehicle was maintained generally at low driving speeds; therefore, approximately 3-4 hours of
driving time and an additional 1 hour stationary sampling period was performed. The stationary
sampling by the mobile vehicle was conducted primarily to compare side-by-side with the stationary
monitoring site. In addition to electric power for driving, the vehicle also had built-in inverters
providing up to 2 kW of power for on-board instrumentation.
The air monitoring instruments utilized for mobile monitoring were selected for a very fast response
time and sensitivity at ambient levels, allowing accurate air quality readings and spatially resolved
data while the vehicle is in motion. Details regarding the instrumentation, driving path, data
processing, and quality assurance are provided below.
2.3.1. Mobile Monitoring Instrumentation
A list of the instrumentation used on-board the mobile sampling platform is provided in Table 2-3.
Gaseous measurements included CO and SO2, both measured in parallel using a quantum cascade
laser instrument. This instrument is able to measure trace gas concentrations with high specificity
and with manufacturer-reported sub-ppb sensitivity while measuring at one second time intervals.
17
-------
As described in section 3.1.1.1, the SO2 measurement ended up being unsuccessful due to poor
performance of the on-board laser and was not able to meet the in-field quality checks. The CO
laser had excellent performance throughout the mobile campaign.
Particle measurements were conducted using optical-based methodologies that provided the
nominal sampling time resolution of 10 seconds or faster. Particle size-resolved number
concentration was measured using two instruments with the detection method optimized in each
instrument for the size regime being sampled. The very small particles - 5.6 to 560 nm in size or
0.0056-0.560 u.m -were measured based on particle electrical mobility using an Engine Exhaust
Particle Sizer (TSI, Inc.). Briefly, the particles are drawn into the instrument, charge-neutralized,
then provided a surface charge. The particles then travel along a column of stacked charged rings,
which attract and then count particles of specific sizes based upon how the charged particle moves
in an electric field. In this way, the instrument is capable of isolating and simultaneously counting
particles of different sizes. The larger particle range - 0.5 - 20 u.m - was measured based upon
inertial principles. The Aerodynamic Particle Sizer (TSI, Inc.) continuously draws in a sample and
then accelerates the sample air flow. The particles accelerate based upon their size and are counted
via light scattering. These two instruments together provide a size-resolved number concentration,
with the size range of 0.0056-20 u.m sampled in 84 discrete size intervals. This size-resolved particle
number concentration can be used to approximate a mass concentration using the following
assumptions - particles are spherical with a diameter at the mid-point of the size bin and particles
have a density of a certain value.
Black carbon (BC) is another particle measurement that was conducted on-board the mobile
platform. BC is the sole measurement that has nearly identical measurement methodology in the
stationary monitoring site (Table 2-2). The instrument operates by drawing a sample air stream
through a filter, with a red light beam (880 nm) passing through the filter and a detector
continuously reading the change in light attenuation over time due to light-absorbing particles
(Ipm) (Hansen et al. 1984). The model used on-board the mobile platform (AE-42) was customized
for higher time-resolution readings by using a relatively small particle deposition area ("spot size")
and a high flow rate (4 Ipm).
18
-------
Table 2-3. Mobile monitoring instrumentation
Measurement
Rate
Instrument
Carbon monoxide (CO)
Sulfur dioxide (SO2)
Particle number concentration
(5.6-560 nm, 32 channels)
Particle number concentration
(0.5-20 urn, 52 channels)
Black carbon
Longitude and latitude
1 s Quantum cascade laser (QCL, Aerodyne Research, Inc.,
Billerica, MA, USA)
1 s Quantum cascade laser (QCL, Aerodyne Research, Inc.,
Billerica, MA, USA)
1 s Engine Exhaust Particle Sizer (EEPS, Model 3090, TSI, Inc.,
Shoreview, MN, USA)
10 s Aerodynamic Particle Sizer (APS, Model 3321, TSI, Inc.,
Shoreview, MN, USA)
1 s Single-channel Aethalometer (Magee Scientific, AE-42,
Berkeley, CA, USA)
1 s Global positioning system (Crescent R100, Hemisphere GPS,
Calgary, Alberta, Canada)
Additional instrumentation on-board the electric vehicle included a high-resolution global
positioning system, which has been tested to have a spatial resolution of <1 meter while sampling at
1 Hz. The global positioning system (GPS) requires overhead satellites to provide positioning
information - lack of satellite detection due to blocked (e.g., dense buildings) or interfering signals
can lead to areas where position data are not known. To understand if the GPS would have any
troublesome areas, preliminary drives were conducted. The key areas of the route were detected
correctly by the GPS; however, when the car was positioned near the stationary monitoring location
the GPS signal dropped off for reasons unknown. Another on-board instrument was a forward-
facing webcam that was used to continuously record the driver's view over the course of each
sampling session.
Finally, during the window of time when mobile monitoring sessions occurred, an ultrasonic
anemometer (RM Young, Model 81000) was positioned on the rooftop of the BNSF office building to
measure 3-dimensional wind speed and direction. This building is located on the northern boundary
of the rail yard (yellow star in Figure 2-2). These data were recorded at a very fast time-rate (4 Hz)
and then were sub-sampled to be 10 s data in order to have more manageable file sizes.
2.3.2. Driving Route
The mobile sampling car driving route was designed to meet several key criteria - 1) ability to repeat
the route at least three times within a given 2-3 hour driving session, allowing areas to be sampled
repeatedly, 2) measuring air quality in low-traffic neighborhoods on the prevailing downwind side of
the yard and areas representative of the urban background, 3) driving safety and minimal time on
rough road surfaces which could give the vehicle hard jolts and affect instrument performance.
The final driving route was selected when staff arrived in the field and test drove possible roadways;
the route is shown in Figure 2-2. For each lap of the route, the vehicle initiated at the southern side
of the rail yard, traveled along the boundary roadway surrounding the yard (going West and then
North), then exited to travel along areas to the North of the yard, followed by crossing over the rail
yard to sample an area to the southeast of the yard, and finally returning to the starting location.
19
-------
Some key areas to note for the route include low-traffic areas designated as urban background
(BGD1 and BGD2), with the assumption that these areas would represent local air quality conditions
without any major sources emitting in the immediate vicinity. BGD1, the northern background area,
was estimated to receive winds from the surrounding area but not from the rail yard during winds
from the west, north, and due east (195-360 degrees and 0-90 degrees). Meanwhile, BGD2, the
southeastern background area, was estimated to receive winds from the surrounding area but not
directly from the rail yard area under conditions of winds from the south and east (75-220 degrees).
Also shown in Figure 2-2 are several neighborhood transects of interest (NT1-NT4), which were
determined to be low traffic residential roadways that spanned from very near the northern side of
the rail yard to several hundred meters in downwind distance.
Additional locations marked include the location of the temporary meteorological monitoring
location during the mobile sampling (yellow star) and the location of the stationary monitoring site
(orange star). After the driving route was completed on any given session, the sampling vehicle
was then parked at the stationary monitoring location which is directly NE of the rail yard.
Figure 2-2. Mobile sampling vehicle driving route surrounding the rail yard of interest. Background (BGD) and
neighborhood transect (NT) areas are noted, as well as the location of the meteorological station put in place
during the mobile sessions (yellow star) and the stationary monitoring site (orange star).
20
-------
2.3.3. Mobile Data Processing
Analysis of mobile air monitoring data includes a series of data-processing steps that carefully time-
align the location and measurement data, apply specific data filtering procedures depending on the
analysis objectives, spatially reference the measurement locations against an area of interest, and
then final analyses are initiated (Figure 2-3). For CIRYS, mobile monitoring data processing involved
first logging the raw data to two onboard computers that were time-synchronized prior to each
drive to the GPS timestamp. One computer logged the data from the EEPS, APS, Aethalometer, GPS,
and webcam. The second computer was built into the quantum cascade laser system and logged CO
and SO2 data.
Raw data
from
instruments
Time-align GPS
& air pollution
data
Apply ONA smoothing
procedure for BC data
Hagleretal.. 2011
J
Data from specific zones of
interest are extracted, distance
referenced to rail yard
Evaluation of downwind areas
(50 m) relative to urban
background
Apply side road traffic exhaust
filter (5-second moving
coefficient of variation for CO)
Figure 2-3. Mobile data processing steps.
After data collection, data processing included providing a time-adjustment of ~l-5 seconds to
account for the time it took for a change in concentration at the sample inlet to be registered on a
particular instrument, which is a function of the flow rate and response time of each instrument. A
change in concentration was provided by applying a zero filter to the inlet for the particle
measurement instruments and by providing a concentration change using a gas standard for the
gas-phase measurement instruments. Following this time adjustment, the air monitoring data were
synchronized with the GPS data, providing an air quality data series that is a function of time and
location.
21
-------
Aethalometer data required an additional data processing step - the application of the Optimized
Noise-Reduction Algorithm (ONA), which has been shown to significantly reduce noise in real-time
BC data while preserving significant trends (Hagler et al. 2011). This instrument operates by
observing the attenuation of red light (880 nm) through a particle-laden filter at the start and end of
each sampling period, translating the difference in light attenuation into a concentration. At low
concentration periods, the addition of new light-absorbing particles to the filter may be insufficient
to change the light attenuation enough to overcome the background noise, leading to a possible
under-prediction of concentration in one instance followed by an over-prediction of concentration
in the next instance (or vice versa). However, over time, the loading of light-absorbing particles
leads to a meaningful decrease in the light signal and translates to an accurate BC concentration
estimate. The ONA algorithm uses the light detection data ('ATN') that is logged by the instrument
to guide the appropriate smoothing of the data. Other commercially available high-time resolution
air monitoring instruments perform similar functions internally and output noise-reduced data-the
Aethalometer is unique in having this necessary smoothing process deferred to the user.
Prior to mobile data analysis occurring regarding near-rail yard air pollution, a final data processing
step is a screening algorithm seeking to detect and remove any instances of local vehicular exhaust
that may provide a bias in the data. The mobile monitoring schedule and driving routes were
designed to minimize the occurrence of local vehicular exhaust. However, unpredictable events may
have occurred and an algorithm seeking short-term spikes in a pollutant indicative of local exhaust
impacts is one objective data-screening approach (Hagler et al. 2010, Hagler et al. 2012). As this
study was seeking to characterize local level impacts due to what was anticipated to be
concentrated diesel emissions within the rail yard environment, CO was selected as the primary
indicator to selectively remove impacts of side road gasoline vehicle exhaust within the time series.
This algorithm performs a running calculation of a 5-point standard deviation in CO (logged in 1
second intervals) and divides this value by the ~2 hr session mean value, providing a modified
running coefficient of variation. The algorithm then flags and removes from analysis all 5-second
time periods where the standard deviation more than doubles the session mean. This procedure
typically flags approximately 1-3% of the 1-second CO data collected. An example of the data
flagging results for CO is provided in Figure 2-4. One should keep in mind that the full time series of
data, as shown in Figure 2-4, also includes portions of the route that are known to have significant
on-road traffic and were utilized to access lower traffic areas that are the primary areas for analysis.
22
-------
2.5
2
1.5
a,
x10
0
1
0.5
21:00
original
- after CO spike detection
22:00
Figure 2-4. Example time series of carbon monoxide data, showing the original raw data (gray) and the data
time series after detected significant short-term spikes were removed, which were presumably due to local
gasoline vehicle exhaust impacting the measurements.
After these preliminary processing steps, the mobile data set for a particular 3-hour session were
now prepared for data analysis and interpretation. For a particular session, data relevant to specific
areas of interest (Figure 2-2) were extracted from the time series based upon the recorded
longitude and latitude. For example, the mobile vehicle may have traveled slowly along the NT1
neighborhood area four or five times during a sampling session, collecting data continuously while
driving the path. The data relevant to the NT1 area would be extracted and then the data could be
grouped into specific distance ranges from the rail yard boundary, with the boundary being
estimated as the northern route driven by the mobile vehicle surrounding the yard (Figure 2-2). For
the analyses shown in this report, the distance between a transect location and the rail yard
boundary was calculated as the shortest distance between the two, with the rail yard boundary
estimated as shown in Figure 2-5. An example of a concentration versus distance from rail yard is
provided in Figure 2-6. When the real-time data were grouped into 50 m intervals, each interval for
neighborhoods NT2-4 generally had thirty 1-second data points per session and NT1 50 m interval
typically had ninety 1-second data points per session; the higher number of points in NT1 is due to
this area having a somewhat U-shaped transect with more measurement time spent per spatial
increment.
23
-------
41.845r
41.844
41.843
D)
41.842
.;£ 41.841
CO
41.84
41.839
41.838
distance (m)
200 300 400
500
rail yard
boundary
shortest distance
between driving
location and
rail yard boundary
-87.773-87.772-87.771 -87.77 -87.769-S7.768-87.767-87.766-87.765
Longitude (deg)
Figure 2-5. Example of distance estimated from the pathway driven in neighborhood NT1 relative to the rail
yard edge. The dashed line demonstrates the shortest distance from an example location along the driving
route to the estimated rail yard boundary line. The shortest distance from each location in neighborhood NT1
to the rail yard boundary is calculated and can be compared to the color bar at the top of the figure.
24
-------
ROD
500
400
s
•S 300
O
O
200
100
(
' ^^^^ - ^^^ I
-
•
•
} 50 100 150 200 250 3C
Distance from rail yard boundary (m)
K)
2500
cn
^,2000
!Z
S 1500
ro
O
o 1000
CD
CQ
500
°C
,
-------
2.4. Stationary Sampling Methods
The stationary data set covers the time period from October 2010 to October 2011. Air monitoring
measurements were taken using monitors typically used for National Ambient Air Quality Standard
(NAAQS) comparisons; these monitors were housed in an air monitoring station located in a lightly
used parking lot to the northeast of the Cicero Rail Yard. The station provided power,
heating/cooling, and a workspace for operators. The deck on the roof of the station provided space
for the PM2.5 and meteorological monitors and the inlets for the gaseous pollutants. The monitors
were connected to a data logger, and the data were downloaded nightly or manually during an
operator site visit.
The air monitoring instruments utilized for the stationary monitoring were borrowed from a number
of agencies both from within and outside of Region 5. The Region attempted to use equipment
similar to what is typically used at National Air Monitoring Stations (NAMS) or State or Local Air
Monitoring Sites (SLAMS). Details regarding instrumentation, data processing, and quality assurance
are provided below.
2.4.1. Instrumentation
A list of the instrumentation used in the station is provided in Table 2-4. Gaseous measurements
included CO, NO, NO2, NOx, and SO2; particulates (PM2.5) and black carbon were also measured.
Meteorological parameters measured included wind speed and direction, relative humidity,
temperature, and barometric pressure.
Data for the gas pollutants were stored in an ESC data logger, and collected nightly via a cell
modem. The data for the Aetholometer were stored on an internal floppy disc; the disc was
collected monthly during a site operator visit. The meteorological data were collected by a
dedicated data logger; these data were downloaded manually each week during a site operator
visit.
Table 2-4 Stationary monitoring instrumentation
Measurement Rate Instrument
Carbon monoxide (CO) 1 min API 200E (Teledyne API, San Diego, CA, USA)
Sulfur dioxide (SO2) 1 min API 101E (Teledyne API, San Diego, CA, USA)
5 min E-BAM Mass Monitor (Met One Instruments, Grants Pass,
PM" /I hr OR, USA)
Oxides of 1mm TECO 42 (Thermo Scientific, Franklin, MA, USA)
Nitrogen(NO,NO2,NOx) v '
Black carbon 5 min AE 21 (Magee Scientific, Berkeley, CA, USA)
. ..... 5 min Cup anemometer (Met One Instruments, Grants Pass, OR,
Ambient wind speed MC/U
\JJr\J
Ambient wind direction 5 min Wind vane (Met One Instruments, Grants Pass, OR, USA)
Ambient relative humidity 5 min RH sensor (Met One Instruments, Grants Pass, OR, USA)
26
-------
2.4.2. Stationary Monitoring Location
The site was located in the parking lot of the Cicero Public Works building, approximately 50 meters
from the nearest train tracks bordering the upper northeast area of the rail yard (Figure 2-7). This
site was selected after surveying a number of possible locations surrounding the rail yard, and met
the desired attributes of being on the prevailing downwind side of the rail yard and being in close
proximity to the rail yard. The southern end of this parking lot borders the BNSF rail yard. The
northern end is bordered by a low volume road (W 26th Street). The area is mostly residential with
some commercial buildings and several industrial sources.
The tracks located approximately 50 m from the station were in routine use throughout the
monitoring study by both commuter and freight trains. This near-field emissions source was
anticipated to contribute short-term concentrated plumes which would be captured in the
monitoring data. However, it was uncertain prior to measurements whether or not the 5 minute
data would be able to isolate time periods affected by train plumes or not. Part of the study design
therefore included periods where the mobile monitoring vehicle would park adjacent to the
stationary monitoring site and collect higher time-resolution measurements for comparison
purposes (section 3.1.1.4).
Figure 2-7. Sampling location, where the stationary monitoring site was located for approximately a year's
time and the mobile vehicle would park for 1-2 hour intercomparison periods.
27
-------
The stationary monitoring data set included several processing steps - data collection either internal
to instruments or using external dataloggers, quality assurance review of the data, consolidation of
the multiple measurements into a single multipollutant time series, and finally temporal and wind-
directional analyses using several approaches. Data for the gas pollutants were stored in an ESC
data logger, and collected nightly via a cell modem. The data from the Aetholometer were stored
on an internal floppy disc; the disc was collected monthly during a site operator visit. The
meteorological data were collected by a dedicated data logger; there data were manually
downloaded each week during a site operator visit.
As discussed in section 3.1.1.4, the near-field effects of passing trains south of the monitoring site
were unidentifiable to be flagged in the 5-minute data time series. Therefore, the stationary
monitoring data set represents general ambient trends at this location, which would be affected by
both the near-field passing trains under certain wind conditions as well as other sources in the area.
Data for the stationary site were reviewed for quality using the criteria listed in Table 2-5 by the
Region's quality assurance coordinator. After the quality-assured data were consolidated into a
single multipollutant time series, analyses were conducted using the R statistical package (R
Development Core Team 2008). The R program, a common platform upon which several custom
functions have been developed, is helpful for air monitoring research. One group of functions used
in this analysis were developed by the OpenAir project (http://www.openair-project.org/), which
provides algorithms to generate wind roses, concentration roses, and time series figures.
Further exploration of the stationary monitoring data trends was accomplished using the
nonparametric trajectory analysis (NTA) model that has been recently developed by EPA's Office of
Research and Development (Henry et al. 2011). NTA is a technique which utilizes highly time
resolved data collected at five minute intervals to identify the possible location of sources in close
proximity to a monitoring site. Backward wind trajectories from the monitoring site are constructed
from the wind data and associated with a single measured concentration at the time when the air
parcel arrives at the monitor. A spatially weighted average then aggregates the pollutant
concentrations so that areas of high concentration are isolated for further study as well as
quantifying possible contributions from potential local sources.
Usually, the five minute averaged wind data are used to create trajectories that go back one hour in
time. The path taken by the trajectory shows the area where the air parcel passed over for the
previous hour at a total of 12 endpoints. The pollutant concentration associated with the most
recent 5 minute measurement is then assigned to each node along the trajectory's path (Figure 2-8).
28
-------
• Time
12/23/20100:55
12/23/20100:50
12/23/20100:45
12/23/20100:40
12/23/20100:35
12/23/20100:30
12/23/20100:25
12/23/20100:20
12/23/20100:15
12/23/20100:10
12/23/20100:05
12/23/20100:00
Wind Direction 1
305.8
306.1
308.1
306.5
303.8
315.4
305.5
307.3
316.9
323.3
319
317.6
2.6 | -
2.3
1.8
2
2
2.2
2 |
2.4 i o -
2.7 5
2.3
2.1
1,5
o -
17(0 2010-12-230000 00
17(0,
1780^
178«x
Iw9
° X
' 71 780 201 0-1 1-13 00 JO 00
"IV
*%,
1740,1010.12-23005500
• Cicero Mgnrtonng Site
-SttOO -4000 -2(100 0 2000 4000 6000
X Coordinate («i««it)
Figure 2-8. Example wind data showing how a typical backward trajectory is constructed for NTA. The twelve
endpoints making up the hour long trajectory are listed in the table at the left and plotted in the graph. The
value of 178 represents a concentration measured at time 2010-12-23 00:55:00.
Repeating the trajectory construction process for all 5 minute measurements yields a large cluster of
trajectories along with their associated 5 minute pollutant concentrations (Figure 2-9). A spatial grid
is laid over the trajectories and a weighted average is computed for each cell by using a kernel
function. The purpose of the kernel function is to weigh the values associated with the nodes in
proximity to the cell's center so that points further away from the center are weighted less than
those closer to the center. The expected concentration at each cell's center is represented by the
equation:
ym yn
m yn
i=l A/=l
(eql)
where
and
K(u) = 0.75(l-u2) for|u|<=l
K(u) = 0 otherwise
(eq2)
(eq3)
(eq4)
The function K shown above (equations 2-4) represents the Epinechikov kernel function where any
points outside the boundary set by the smoothing parameter (h) have a zero weight and, therefore,
are not used in calculating the average for the cell. The smoothing parameter represents the
29
-------
distance from the center of the cell of the endpoints to include in the weighted average. Varying
this parameter controls the smoothness of the overall concentration field. A smaller value will make
the field appear more "spiky" while a larger value will encompass more points and make the field
appear more smoothed.
o
o
o
o
o
CM
O
O
U o
>_ o
> o
o
I
o
o
o
o
o
n
-20000 -10000 0 10000
X Coordinate (meters)
20000
30000
Figure 2-9. Example showing all of the backward trajectories for all 5 minute averaged data collected at the
Cicero stationary monitor. The red box represents a grid cell for calculating the weighted average at the cell's
center. The cell's expected concentration would be calculated based on the distance individual trajectory
endpoints are from the grid cell's center.
The results of the pollutant spatial fields are displayed on Google Earth maps to determine possible
local sources that could be contributing to the short term measured concentrations at the Cicero
stationary monitoring site. The NTA expected concentrations were calculated for black carbon, SO2
and NO for a period from November 2010 through May 2011 when all three pollutants were being
measured concurrently. The time period encompasses the full measurement record for black
carbon and includes, for comparability purposes, the corresponding SO2 and NO measurements.
Due to the diverse local source mix within the vicinity of the Cicero rail yard, a grid spanning 20 km
in both the north-south and east-west directions was created and centered at the monitoring
location just north of the rail yard boundary. The grid's extent included other possible source
30
-------
contributors such as the Crawford power plant, Midway Airport, the McCook industrial zone to the
southwest as well as other rail yards within the general vicinity. These results are discussed in
section 3.2.1.2.
2.4.4. Quality Assurance
Quality Assurance was an important consideration for the stationary monitoring effort. To that end,
the Region wrote a Quality Assurance Project Plan, which outlined the process and targets to assure
the data was of sufficient quality to meet the objectives of the study. Methods, frequencies, and
criteria goals for the stationary monitoring are summarized in Table 2-5 below.
Table 2-5. Stationary monitoring site quality assurance procedures
Measurement
Parameter
PM2.5
Black Carbon
NO,NO2,NOX
S02
CO
Ambient wind
speed and
direction
Assessment Method
Flow Checks; flow check
using independent standard
Flow checks
Single point QC check; Audit
(consecutive levels at 80% of
measured cone.)
Single point QC check; Audit
(consecutive levels at 80% of
measured cone.)
Single point QC check; Audit
(consecutive levels at 80% of
measured cone.)
Certification, general
inspection and maintenance
Minimum Frequency
Monthly; every 6
months
Monthly
Every 2 weeks;
Every 6 months
Every 2 weeks;
Every 6 months
Every 2 weeks;
Every 6 months
Pre-deployment;
Every 6 months
Criteria
±4%
±4%
-------
3. Data Analysis
3.1.Mobile Sampling
3.1.1. Mobile Data Overview
A total of 23 mobile sampling sessions were performed between October 27, 2010 and
November 21, 2010 and are listed in Table 3-1. Each mapping period was conducted for
approximately 3-4 hours, followed by parking the sampling vehicle next to the stationary air
monitoring shelter for a period of 1-2 hours to allow for intercomparison between
measurements. The total sampling duration was constrained by available power in the electric
vehicle.
Table 3-1. Mobile Sampling Sessions
Date
(start)
27-Oct-10
28-Oct-10
29-Oct-10
30-Oct-10
31-Oct-10
l-Nov-10
3-Nov-lO
4-Nov-10
5-Nov-lO
6-Nov-10
7-Nov-10
8-Nov-lO
10-Nov-lO
ll-Nov-10
12-Nov-10
13-Nov-lO
15-Nov-lO
16-Nov-10
17-Nov-10
18-Nov-lO
19-Nov-10
20-Nov-10
21-Nov-10
Time
category
Mid-Day
Evening
Evening
Mid-Day
Early
Morning
Evening
Mid-Day
Early
Morning
Mid-Day
Early
Morning
Evening
Evening
Mid-Day
Early
Morning
Mid-Day
Early
Morning
Evening
Evening
Mid-Day
Early
Morning
Early
Morning
Early
Morning
Evening
Day of week
Wednesday
Thursday
Friday
Saturday
Sunday
Monday
Wednesday
Thursday
Friday
Saturday
Sunday
Monday
Wednesday
Thursday
Friday
Saturday
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Sunday
Driving mode
Start End Duration
9:16
18:53
18:45
8:52
3:52
19:18
11:50
4:10
9:00
3:52
19:40
19:00
9:10
4:00
10:00
4:00
19:30
18:55
9:45
3:58
3:52
3:57
19:02
13:05
0:50
22:45
12:15
7:23
23:22
15:25
7:45
12:23
7:15
23:09
22:52
12:30
7:10
13:41
7:00
23:10
23:30
12:52
7:00
7:08
8:06
22:30
3:49
5:57
4:00
3:23
3:31
4:04
3:35
3:35
3:23
3:23
3:29
3:52
3:20
3:10
3:41
3:00
3:40
4:35
3:07
3:02
3:16
4:09
3:28
Stationary mode
Start End Duration
13:05
0:55
22:45
12:15
7:25
23:22
15:25
7:45
13:50
7:15
23:09
22:52
12:30
7:10
13:41
7:00
23:10
23:30
12:52
7:00
7:08
8:06
22:30
13:25
2:00
23:45
13:15
8:25
0:10
16:25
8:42
14:30
8:15
23:56
0:10
14:00
9:40
15:05
8:40
1:05
1:30
14:37
8:42
8:30
9:49
0:09
0:20
1:05
1:00
1:00
1:00
0:48
1:00
0:57
0:40
1:00
0:47
1:18
1:30
2:30
1:24
1:40
1:55
2:00
1:45
1:42
1:22
1:43
1:39
32
-------
The sampling vehicle repeatedly followed a driving route that surrounded the rail yard (Figure 3-2),
but emphasized capturing data in areas to the northwest of the rail yard - area downwind of
prevailing wind. The vehicle traveled along roadways that immediately bordered the rail yard -
Ogden Ave to the south, an interior roadway to the rail yard, and W 26th St to the North. The vehicle
covered the path shown approximately four times per sampling session. The vehicle was driven as
slowly as possible to optimize the spatial resolution of the data - the majority of the route was at
rates of 5-25 mph (Figure 3-1). Images of the sampling vehicle taken during the field study are
provided in Figure 3-2.
Speed
20 (mph)
41.838
41.836-
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
Longitude
Figure 3-1. Driving speed recorded by the sampling vehicle's GPS
Figure 3-2. Picture of the instrumentation in the vehicle (left) and sampling vehicle in action (right) at the
Cicero Rail Yard.
33
-------
As detailed in the Quality Assurance Project Plans (QAPP - Appendix A), pre-study checks and daily
quality checks were conducted to verify the performance of air monitoring instruments. The pre-
study checks included flow rate verification for all analyzers. The daily checks included:
• Gas-phase species (carbon monoxide, sulfur dioxide): within +/- 20% of a calibration gas
• Particulate matter measurements (black carbon, fine/coarse particulate matter, ultrafine
particle counts): zero check by applying a high-efficiency particulate air (HEPA) capsule filter
to the inlet and observing concentrations for several minutes.
QC checks were performed both before and after each sampling drive. An example analysis of QC
check results is shown in Table 3-2 and the full set of QC checks is shown in Appendix B. As shown,
CO checks were within 2% of the gas standard, whereas SO2 failed to meet the quality objective of
within 20%. While these two measurements were collected using the same instrument platform
(dual quantum cascade laser), they were measured using independent lasers and the SO2 laser
performance appeared to have been affected by temperature control issues. The failure of the SO2
laser was detected during the mobile monitoring field intensive and correspondence with the
manufacturer occurred to diagnose the cause of the instrument failure. After several attempts to
troubleshoot the laser during the study, it was determined that in-field repair of the SO2 laser
system was not possible and also could not remotely occur in a timely fashion. Considering the
value of other pollutant measurements occurring and project budget/schedule constraints, the field
study continued and due to failed quality checks, the SO2 data were discarded for data analysis and
will not be evaluated in this report. The CO data met data quality objectives and had 100%
completeness (Table 3-3).
The daily zero checks on the particulate matter instruments served to determine whether or not the
connections between the instrument and the inlet were airtight, as well as to indicate whether the
instruments were responsive to a sudden and significant change in particulate concentrations. The
HEPA filter is rated to remove 99.97% of particles above 0.3 u.m in diameter. It is possible that
smaller particles (<0.3 u.m in diameter) have a higher penetration through the filter to the
instrument. Therefore, the zero check was estimated by looking at particle counts for particles
greater than 0.1 um in size for the EEPS. Black carbon particles, also often smaller than 0.3 u.m, may
have penetrated through the filter and caused non-zero readings during the zero check; however,
the readings were still within the target of <20% of ambient (Table 3-2). The zero checks were
calculated by averaging the time period of data collected when the filter was in place on the inlet
and dividing this value by the ambient concentration measured over the 30 min either before or
after the zero check, depending on whether the check occurred at the start or end of the daily
sampling session. An example time series showing a zero check is provided in Figure 3-3. Overall,
the particulate measurements had nearly 100% completeness, with only one instrument data
flagged and removed for analysis (Aethalometer) for a single sampling session.
34
-------
Table 3-2. Example QC metrics for the air monitoring instruments onboard the sampling vehicle
Sampling Aeth-Zero
Date (filtered/
ambient * 100%)
11/3/2010
11/4/2010
11/5/2010
2%
9%
7%
EEPS-Zero
(filtered/
ambient * 100%)
1%
0%
0%
APS-Zero
(HEPA-filtered/
ambient * 100%)
0%
0%
0%
CO
(measured/cal*
100%)
99%
98%
98%
S02
(measured/cal
*100%)
69%
67%
68%
Table 3-3. Data completeness
Measurement
Completeness
(#/23 sessions)
Quantum Cascade Laser - Carbon monoxide
Quantum Cascade Laser - Sulfur dioxide
Aethalometer - Black carbon
Aerodynamic Particle Sizer- Fine/coarse particle counts
Engine Exhaust Particle Sizer - Ultrafine particle counts
Global positioning system - Longitude and latitude
Ultrasonic anemometer - Wind speed, wind direction, temperature
100%
0%
96%
100%
100%
100%
100%
,x10
E
O
m 0.5
0
"E is
1
"10
-
I
zero filter on
!_ Js
i
LL^___J
i i
19:00
19:10
19:20
*** . *** * + » '**** ** **(.* ***
MV»*VV~..' • -.'.•••%. "•...«.-*
19:00
19:10
19:20
19:30
19:30
19:30
Figure 3-3. Example particulate time series during the addition of a zero filter and after the filter is removed.
The filter-on period is marked with the light blue box.
Additional measurements used in the analysis include the global positioning system to determine
real-time location as well as an ultrasonic anemometer to measure meteorological parameters. As
discussed in the QAPP, the GPS data were verified against the driver's observations (e.g., start time
35
-------
and stop time from a particular location) and the planned driving route. The wind sensors were
aligned properly and compared against operator observations (e.g., observation of strong winds
from the South matching measurements).
A period of wind data collected on the rooftop (24 hr period from 10/26/2010 15:00-10/27/2010
15:00) was compared with a nearby meteorological station at Midway Airport, located
approximately 4 miles south of the rail yard. The wind data at Midway Airport were collected at 1.5
m above ground at hourly intervals, while the study meteorological sensors collected data on top of
a 3-story building (~12 m above ground). As shown in Figure 3-4, both sensors agreed with the
general range of wind speed and orientation of wind direction. However, the much more rapid
sampling of the sensor used in this study (10 s data), provided approximately 8600 data points
versus 24 data points collected at Mid-way, which explains the better defined wind rose in Figure 3-
4.
Midway Airport (hourly data)
-,,.,_ SOUTH,--'
wind
speed
Rooftop ultrasonic anemometer data (2 s data)
-—..SOUTH-"
Figure 3-4. Comparison of wind measurements collected at Midway airport (left) with wind data collected on
top of a BNSF building (right).
Another comparison was made between the ultrasonic anemometer data collected during this study
and data from the Region 5 air monitoring station, situated near the northwest corner of the rail
yard; the Region's sensors were placed 2 m above the ground. The Region 5 station collected wind
speed and wind direction data at 5 min increments. Wind roses were generated for a 24 hour time
period (11/17/2010 00:00 - 11/18/2010 00:00) using each station's data. As shown in Figure 3-5, it
can be seen that the very high data rate collected using the ultrasonic anemometer yielded a
broader distribution of observations. However, both sets of data are in agreement in terms of
general wind direction and wind speed.
36
-------
Region 5 Station (5 min data)
/'
/ /
A'EST — i — — u —
~"~--.^
v.
20% \
10% >,
asr
/ B14-16
.-'*" / ," a 12- 14
a 10- 12
ns-io
.-''"' / D6-8
CH4-6
D2-4
SOUTH-''''' D°-2
Rooftop ultrasonic anemometer data {2 s data)
•*'
/ :#4\
~- .. __
'•--"- — t
'•Urs.! rn ~ •• .
••.
/ B14-16
---'' / ,' O12-14
D10-12
DS-10
..--''' ,' DS-8
02-4
SOUTH--''' DO-2
Figure 3-5. Comparison of wind measurements collected at the stationary monitoring location (left) with wind
data collected on top of a BNSF building (right).
3.1.1.2. Sampling sessions in the context of meteorology
Local meteorology and rail yard emissions activity are anticipated to be critical factors driving the
degree of local air quality impact associated with proximity to a rail yard. Average wind speed and
wind direction were calculated for each mobile sampling session, from the point when the vehicle
began driving to when stationary sampling was completed (Table 3-4). Wind roses were generated
on 15 min intervals and for the entire session for each sampling time frame (Appendix C). Wind
direction standard deviation was calculated using equations (5)-(6), below (Yamartino 1984).
_ _
e -
A vn
-------
Table 3-4. Wind characteristics during mobile sampling sessions
Session
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Start time
10/27/20109:16
10/28/2010 18:53
10/29/2010 18:45
10/30/2010 8:52
10/31/2010 3:52
11/1/2010 19:18
11/3/2010 11:50
11/4/20104:10
11/5/2010 9:00
11/6/20103:52
11/7/2010 19:40
11/8/2010 19:00
11/10/20109:10
11/11/20104:00
11/12/2010 10:00
11/13/20104:00
11/15/2010 19:30
11/16/2010 18:55
11/17/2010 9:45
11/18/2010 3:58
11/19/20103:52
11/20/2010 3:57
11/21/2010 19:02
End time
10/27/2010 13:05
10/29/2010 2:00
10/29/201023:45
10/30/2010 13:15
10/31/2010 8:25
11/2/2010 0:10
11/3/2010 16:25
11/4/2010 8:42
11/5/2010 14:30
11/6/2010 8:15
11/7/2010 23:56
11/9/2010 0:10
11/10/2010 14:00
11/11/2010 9:40
11/12/2010 15:05
11/13/2010 8:40
11/16/2010 1:05
11/17/2010 1:30
11/17/2010 14:37
11/18/2010 8:42
11/19/2010 8:30
11/20/2010 9:49
11/22/2010 0:09
U, scalar
(m/s)
8.0
3.1
4.6
6.4
2.4
1.9
4.0
4.1
5.1
1.8
1.9
2.1
3.6
2.2
3.1
2.8
2.7
2.0
2.9
2.7
3.7
2.2
8.4
U, vector
(m/s)
7.4
2.5
4.5
5.7
2.2
1.7
3.6
3.9
4.7
1.7
1.7
2.0
3.3
2.1
2.8
2.6
2.5
1.9
2.3
2.6
3.5
2.1
8.3
ea(d
eg)
236
294
208
232
352
70
230
327
338
309
184
175
164
184
44
137
192
304
280
321
194
358
212
Oeb
(deg)
33
41
9
32
18
13
29
14
22
17
11
16
25
21
24
21
17
21
48
16
21
19
9
Category
sw
NW
SW
SW
N
NE
SW
NW
NW
NW
S
S
SE
S
NE
SE
S
NW
W
NW
S
N
SW
Vector mean wind direction
Standard deviation of the wind direction, calculated according to eq. (5).
Mean wind speeds ranged from moderate (1-2 m/s) to high (5-8 m/s) throughout the course of
sampling. No sessions were considered to have calm conditions, indicated by low mean wind
speeds (<1 m/s). Wind direction standard deviation is a useful parameter to indicate the degree of
change in wind direction over the course of sampling. Of the twenty-three sampling sessions,
nineteen of the sessions had very low variability in wind direction (<30 deg) while the remaining four
sessions had moderate variability (30-48 deg).
Visuals of mean wind speed and direction per session, organized by sampling time of day, are
provided in Figures 3-6 (early morning sessions, Figure 3-7 (mid-day sessions), and Figure 3-8
(evening sessions). The sampling route had been set to cover the neighboring residential areas
located to the North of the rail yard thoroughly, based on historical wind data indicating prevailing
wind direction from the southwest. Of the cases sampled, approximately half of the sessions had
winds from the SW/S/SE.
38
-------
N
w
Figure 3-6. Morning session wind trends - arrow orientation indicates wind direction (e.g., pointing towards N
means wind from the S) and extent indicates mean wind speed. Sessions shown are #5, 8,10,14,16, 20, 21,
and 22.
N
Figure 3-7. Mid-day session wind trends - arrow orientation indicates wind direction (e.g., pointing towards N
means wind from the S) and extent indicates mean wind speed. Sessions shown are 1, 4, 7, 9,13,15, and 19.
39
-------
N
Figure 3-8. Evening session wind trends - arrow orientation indicates wind direction (e.g., pointing towards N
means wind from the S) and extent indicates mean wind speed. Sessions shown are #2, 3, 6,11,12,17,18,
and 23.
3.1.1.3. Sampling sessions in the context of rail yard activity
The BNSF staff at the Cicero Rail Yard arranged for automatic reports of the minute-by-minute
container lifts ("Lift counts") and trucks passing through the gate ("Gate counts") to be sent daily by
email to EPA Office of Research and Development. These daily data files were concatenated into a
longer time series and used as the basis for calculating hourly and daily totals. Given that the
mobile sampling sessions represented only a several hour window of time on any given day, it is of
interest to understand the rail yard activity both by day of the week and time of day. Figure 3-9
shows the daily lift and gate counts associated with the sampling window (October 27, 2010 to
November 22, 2010). In addition, Figure 3-10 shows the average diurnal rail yard activity trends for
each day of the week over the course of the study period.
40
-------
2000
1800
1600
1400
1 1200
-i— «
A
A
A & A
-
-
-
i i i i i r i
Sun Mon Tue Wed Thu Fri Sat
Gate count 10-2 7-2010 to 11-22-2010
I I I I I L I
*
- ° ° § o -
8 -
n 0
-
-
i i i i i r i
Sun Mon Tue Wed Thu Fri Sat
Figure 3-9. Daily total crane container lifts (a) and truck counts at the gate (b) during the mobile sampling
period of October 27, 2010 through November, 22, 2010.
41
-------
f »
o
i »
o
•50
Sun
Won
•Wed
20 23
I
o
Thu
8 12 16 20 23
•50
•CO
=:
8 12 16 20 23
Sat
12
Hour
• Sun
20 23
• 'A'ed
•Thu
.Fri
• Sat
8 '2
Hcur
20 23
20 23
Figure 3-10. Average gate counts (left) and lift counts (right) by day of week and hour of day during the study
period of October 27, 2010 through November, 22, 2010.
120
100
O
0 60
0)
0
40
20
0
8
12
Hour
16
20
23
Figure 3-11. Comparison of mean diurnal gate activity during the study (heavier dashed black line) with
monthly average diurnal trends from October 2010 to July 2011 (10 months) shown in thin colored lines.
42
-------
Figure 3-12. Comparison of mean diurnal lift activity during the study (heavier dashed black line) with
monthly average diurnal trends from October 2010 to July 2011 (10 months) shown in thin colored lines.
3.1.1.4. Mobile /Stationary data comparison
During the course of the mobile sampling campaign, the mobile sampling vehicle would routinely
park for an hour at the stationary monitoring location and continue sampling with all air monitoring
instrumentation. With the stationary sampling shelter initiating data collection in mid-November,
several sessions towards the end of the mobile sampling campaign are candidates for inter-
comparison of the two data sets. The sampling location is shown in Figure 2-6 and is estimated at
approximately 50 m from the nearest set of train tracks.
An important caveat of this comparison is that the mobile and stationary instruments pulled
samples from different locations both horizontally and vertically and therefore may experience
different concentrations for narrow plumes advected from the near-field. In addition, the
measurements that overlap between the two data series- BC, CO, and PM2.5 -were not sampled
using identical sampling techniques. As the mobile vehicle requires high time-resolution sampling
(1-10 second data) and portable instrumentation, BC was collected using a high-precision portable
Aethalometer, CO was measured using a quantum cascade laser system, and PM2.5 is estimated
from particle count data collected using an Aerodynamic Particle Sizer. Meanwhile, the stationary
site measured concentrations at slower timeframes (5 minutes) - BC was measured using a
rackmount Aethalometer, CO was measured using an NDIR-type rackmount analyzer (FRM), and
PM2.s was measured by beta attenuation (FEM) with a size-selective inlet. Despite these caveats,
the comparison may still yield helpful information in addressing the questions below:
43
-------
• Comparing the real-time data (seconds) collected onboard the mobile car with the 5-minute
data resolution of the stationary monitoring station, are the 5-minute data of sufficient
resolution to isolate periods affected by nearby train emissions?
• How do like-measurements generally compare?
To evaluate these questions, a 5-day series of sampling sessions where the mobile and stationary
data sets overlapped were assessed (11/17 - 11/21/2012) for a general comparison between the
measurements. Within this window of time, the sampling sessions with wind flow from the South -
advecting air from the rail yard area - were studied to determine how near-field train emissions
were captured in the high time-resolution versus 5-minute data series. The comparison of each
sampling session is shown in Appendix B. Shown here is a case study of 11/19/2011 7:00-9:00 AM,
selected due to being a period with winds advecting rail yard / passing train emissions northward
towards the sampling station (Figure 3-13), as well as being a period of time with anticipated
frequent train traffic. The parallel time series of data for BC, CO, and PM2.s is shown in Figure 3-14.
Timespan: 11/19/2010 7:00:00 AM to 11/19/2010 9:00:00 AM
NORTH
30%
20% •.
10% •.
SOUTH
m/s
• 14-16
EH 1 2 - 1 4
IZI10-12
LH6-8
04-6
02-4
LHQ-2
Figure 3-13. Wind rose during a period of mobile and stationary site side-by-side sampling.
44
-------
6
'E 4
en
0 2
CQ
1000
S 500
O
o
80
•T" eo
^ 40
— 20
2 o
x1Q"
-
10 07:20
I
Mobile car * Stationary
^^^A^/vv/^^^^y^^^M^
i
10 07:20
I I I
07:30 07:40
07:50 08
~~\
I
•Wv^ -""*• ,-^*^^Avr^V^^^^V'%^J^
t I I
07:30 07:40
1 | |
„ Mobile car Stationary-5 mm stationary-hourly
-
-
10 07:20
, j r-f^-__^_^ _~ __ ,-J^-Vj*- .-
I I
07:30 07:40
07:50 08
I
07:50 08
rt^^
00 08:10
L~-
00 08:10
-
00 08:10
Figure 3-14. Parallel time series of concentrations for the stationary monitoring site (green) reporting raw data
at 5- minute intervals and the real-time data collected onboard the mobile monitoring vehicle (blue). Note:
Lower limit of detection for the stationary CO analyzer is 300 ppb.
Of the overlapping measurements for the mobile and stationary data sets (CO, PM, and BC), the
real-time BC data appear to have the highest sensitivity to fresh exhaust impacts presumably related
to the upwind rail / rail yard sources. For example, at approximately 0800 (Figure 3-14), BC
increases by a factor of ~45. Meanwhile, for this same episode, CO and PM 2.5 increase by only a
factor of ~2, which is related to the relative abundance of these species in fresh emissions as well as
existing background levels. In addition, for this same spike, UFPs measured on the mobile platform
increase by a factor of ~11. Therefore, of the suite of continuous measurements collected at the
stationary site that overlap with the mobile vehicle measurements, BC has the highest likelihood of
success as an indicator of train events. However, as shown in Figure 3-15, the very short duration of
these spikes - approximately 10 seconds - translate into only a marginal change in concentration
when the real-time data are averaged to five minutes.
45
-------
Mobile car
Mobile car- 5 min average
Stationary
07:30
07:40
07:50
08:00
08:10
Figure 3-15. Black carbon time series for the mobile car (real-time in dark blue, 5-minute average in light blue)
and 5-minute stationary data (green).
Regarding the comparison of the two sets of measurements for BC, PM2.5, and CO, the time series
shown in Figure 3-14 and Appendix B indicate that BC data generally track closely when averaged at
the same time rate; however the comparison for the other two species is challenged by differences
in instrumentation. In the example shown in Figure 3-15, BC concentrations for this particular hour
were generally higher measured at the mobile vehicle location relative to the stationary site - one
possible reason for this is the difference in vertical and horizontal sampling locations for the two
monitors relative to a near-field emissions plume. For PM2.s, the beta attenuation measurement
approach used at the stationary site requires sufficient particle loading over time to reach a
sufficient signal to noise ratio. As shown in Figure 3-14, at a 5-minute averaging rate the PM2.5 BAM
signal oscillates above and below zero, while at an hourly averaging period the signal appears to
settle on a positive value. However, as shown in Appendix B, the hourly data still had negative
values for one intercomparison period (11/20), indicating even longer averaging of those data is
required at lower concentration periods. Given that the intercomparison period between the
mobile vehicle and the stationary site were only in increments of 1-2 hours, the ability to compare
the two PM data streams is limited. However the hourly data and the mobile vehicle Aerodynamic
Particle Sizer data are within a similar general range of concentrations. For the comparison of the
CO data, the CO analyzer onboard the electric car has a sub-ppb level detection limit, whereas the
CO analyzer at the stationary site has a 300 ppb lower detection limit. For most of the comparisons,
the CO stationary site analyzer was recording at or below the limit of detection, which generally
agreed with the concentration range observed by the Quantum Cascade Laser onboard the mobile
vehicle for these time periods.
3.1.2. Assessment of local air quality impact through mobile monitoring
For the purposes of this report, mobile monitoring data were analyzed to address several specific
questions relating to near-source air quality - (1) Are air pollution levels in residential areas
downwind of the rail yard area significantly higher than a similar environment that is upwind of the
46
-------
rail yard? (2) If so, how does this impact vary with wind speed, time of day, and distance from the
rail yard?
For the mobile monitoring analyses to follow, the term "rail yard area" is defined as any emissions
from within the rail yard boundary, which may include commuter and freight trains, switcher
locomotives, truck emissions, and other emissions within the yard area, as well as potential
boundary traffic on two roads immediately adjacent to the yard. One roadway passes immediately
south of the yard (Ogden Rd., 22,000 vehicles, annual average daily traffic) and also lies between the
urban background area to the south and the residential neighborhoods to the north. Another
roadway (26th Street, 9600 vehicles, annual average daily traffic) passes on the immediate north side
of the rail yard between the rail yard and three of the four neighborhoods of interest (NT2-NT4).
These roadways are generally low in terms of traffic volume - 150,000 vehicles annual average daily
traffic is the typical lower threshold for a major highway from the current perspective of local air
pollution impact - however, traffic along these roadways may contribute to the measured
upwind/downwind signals, particularly during higher traffic portions of the day. The fraction of
traffic along these roads associated with the Cicero rail yard is unknown at this time.
With the four neighborhood transects (NT1, NT2, NT3, NT4) shown in Figure 2-2 located on the
northern side of the rail yard, time periods of wind from the south would be considered as
downwind of the rail yard for those areas and periods of wind from the north would be considered
as upwind of the rail yard. The sampling sessions were first organized by time of day into the three
deployment periods - early morning (~3-7 AM), late morning to afternoon (~10 AM - 1 PM), and
evening (~7-10 PM) - followed by categorization by wind trends. This organization allows both
diurnal and wind direction effects to be studied. One important diurnal trend that can affect local
air quality includes the atmospheric mixing height, whereby the solar heating of the earth's surface
during the daytime can enhance upward mixing of air and lower air pollution levels and the reverse
is true during cooler late evening and early morning periods. A second important diurnal trend is
source emissions activity - both sources internal to the rail yard as shown in Figures 3-10, 3-11, and
3-12, and sources external to the rail yard (e.g., street traffic).
Early Morning Periods
Three mobile sessions (#14, #16, #21) captured trends during winds from the south and five mobile
sessions (#5, #8, #10, #20, and #22) captured trends during winds from the north during the very
early morning time period (4-7 AM). An example of a downwind period is shown in Figure 3-16,
below. The primary focus of this analysis is to determine whether there is a statistically significant
exceedance in downwind air pollution levels relative to the general urban background. Therefore,
Figure 3-16 shows specifically the net concentration difference between the downwind measured
(in 50 m spatial increments) and background levels for areas where the comparison between
downwind measurements and the background were determined to be significantly different (refer
to section 2.3.3). Areas outside of these specific neighborhood transects, calculated at distances up
to 300 m from the estimated rail yard boundary, and areas without statistically significant
differences in downwind concentrations are left as black points along the mobile route.
47
-------
The full series of mobile driving maps for each session is provided in Appendix D. Each map can be
viewed to understand both the location and relative concentration difference for a particular
pollutant. As shown in the example session (Figure 3-16), black carbon appears to be the most
sensitive indicator of downwind impact and has exceedances over background levels along all four
neighborhood transects, which is in-line with the observations of diesel emission plumes discussed
in section 3.1.1.4. Meanwhile, carbon monoxide and PM2.5 do not show any significant difference in
comparison to the background for this specific case, and PMi0 and UFPs show only several isolated
areas with an exceedance.
To summarize the broad trends among the eight sampling sessions (three downwind, five upwind),
the four neighborhood areas were considered as twenty-four 50 m segments (0-50, 50-100, 100-
150, 150-200, 200-250, and 250-300 m) and for each species the fraction of segments with
exceedances was calculated. This summary for the early morning period is provided in Figure 3-17.
During the very early morning period, BC appears to have the strongest upwind/downwind signal of
the various pollutants measured, with nearly three-quarters of the downwind areas indicating
statistically significant excess BC levels above the background. For the three sessions measured, this
translated to an absolute excess concentration of 0.3-0.6 u.g m"3 BC. This excess translated to the
downwind neighborhood areas having 30-40% higher total BC concentrations relative to the urban
background (background ranged from 0.8-2.0 u.g m"3 BC). The other measurements shown - UFPs,
CO, PM2.5, and PM10 - do not show the same upwind/downwind trend of excess levels. It is
important to emphasize that this analysis is requiring that the concentration difference meet the
test of statistical significance -one limitation to point out regarding the PM2.s and PMi0 data is that
both were collected at slower time intervals (10 seconds) than the other measurements, therefore
the lower number of data points per area weakens the spatial statistical strength of those two data
sets. This limitation may be partially overcome by grouping the neighborhood transect PM data into
one concentration versus distance evaluation, but these further analyses are not currently explored
in this report.
48
-------
Timespan: 11/11/2010 04:00:00 to 11/11/2010 07:10:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m
-87.79 -87.78 -87.77 -87.76 -87.75
PM25(ngm'J)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
1100
1000
900
800
700
600
500
400
300
_. I" j 200
n
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76
-87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafne Particles (p cm"J)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
PM10(ngm'3)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
110500
110000
§9500
19000
18500
18000
17500
17000
Jesoo
Jeooo
^5500
3.8
3.6
3.4
3.2
3
2.8
2.6
Figure 3-16. Statistically significant excess concentrations above the background for neighborhood transects NT1-
NT4, calculated for areas up to 300 m from the rail yard.
49
-------
ox ,
%of northern
transect area
with significant
increase over
background
BC
UFPs
CO
PM2.5
PM10
Figure 3-17. Fraction of neighborhood areas (NT1-NT4), in 50 m increments up to 300 m from the rail yard, with
significant increase in pollutant levels above the background during early morning sessions (~4-7 AM).
Mid-day Periods
Five sampling sessions were conducted during the late morning to early afternoon and had a clear
upwind (two sessions) or downwind (three sessions) meteorological trend. Of these five sessions, one
upwind session was excluded from analysis as the operator was able to accomplish only two driving laps
(instead of the usual four laps) due to instrumentation troubleshooting. During these sampling periods,
the vehicle operator noted "very high" side road traffic including school buses during this time frame.
As observed in Figure 3-18, during periods of wind from the South, the northern neighborhoods
experienced a broad (75% of the 50 m increments) increase in BC relative to background areas, however
other pollutants measured had sporadic rather than consistent increases above the background. Excess
BC levels ranged from 0.3-1.2 u.g m"3, equivalent to a 62-101% increase over the background (the
background ranged from 0.3-2.0 u.g m"3). During this time period, it is anticipated that the rail yard
would have higher emissions according to diurnal activity data (Figure 3-11 and Figure 3-12) but is also a
timeframe of greater atmospheric mixing. In addition, what is unique to this mid-day time period is that
under periods of wind from the North, these same neighborhoods to the north of the rail yard indicated
elevated BC levels relative to the background for approximately 45% of the neighborhood areas. This
indicates an emissions source located north of the neighborhoods, likely heavier side road traffic during
this time of day. Given that one roadway passes south of NT2-NT4 (Figure 3-2), these results suggest
that side road traffic may contribute to excess BC levels in addition to rail yard emissions during periods
of wind from the South, and the contribution from the rail yard cannot be as distinctly isolated in the
mid-day timeframe.
50
-------
100%
90%
80%
70%
% of northern
transect area 60%
with significant 50%
increase over 40%
background QQ
-------
%of northern
100%
90%
80%
70%
60%
transect area
with significant 50%
increase over 40%
background
20%
10%
0%
BC
UFPs
CO
PM2.5
PM10
Figure 3-19. Fraction of neighborhood areas (NT1-NT4), in 50 m increments up to 300 m from the rail yard,
with significant increase in pollutant levels above the background during evening time periods (7-10 PM).
3.1.2.2. Wind speed effect
The mobile monitoring sessions that were conducted during the early morning and evening periods
- where a clear upwind/downwind trend was detectable for black carbon, were further analyzed to
understand how downwind excess BC compares with local wind speed. For both the early morning
and evening period, a negative relationship was observed between downwind excess BC and wind
speed (Figure 3-20). This relationship indicates that higher wind speeds favor improved dispersion
of local black carbon emissions from the rail yard area and lower the near-source air quality impact.
This result is similar to that determined in near-road field studies for directly-emitted pollutants
(Hitchins et al. 2000, Zhu et al. 2002).
52
-------
Early Morning Sessions Evening sessions
Mean Excess
BC over
background
(ng/m3)
1400
1200
1000
800
400
200
0
- -
0.0 2.0 4.0 6.0 8.0
scalar wind speed (m/s)
10.0
Figure 3-20. Mean downwind BC concentrations in neighborhood areas up to 300 m from the rail yard, as a
function of local wind speed.
3.1.2.3. Impact as function of distance
For the early morning and evening sessions, downwind excess BC concentrations were plotted as a
function of perpendicular distance from the estimated boundary of the rail yard, in 50 meter
increments. In contrast to what has been observed in many near-road monitoring studies
(summarized in Karner et al. 2010), the excess BC concentrations do not follow an exponential
decrease in concentration with distance from the road (Figures 3-21 and 3-22).
These results imply that the spatial extent of downwind impact, in terms of excess black carbon,
likely exceeds 300 m in distance. One influential factor affecting the spatial extent and variability of
the excess BC levels is the densely built environment surrounding the rail yard, which likely affect
the downwind dispersion of emissions. It has been observed in roadway modeling and
measurement studies that even a thin roadside noise barrier can dramatically alter the vertical and
horizontal concentration field in the near-road environment (e.g., Baldauf et al. 2008b).
53
-------
^EarlyAM-1
Early-AM-2
•EarlyAM-3
Normalized
Excess BC
1.20
1.00
0.80
0.60
0.40
0.20
0.00
0 100 200 300
estimated distance from rail yard boundary (m)
Figure 3-21. Normalized downwind excess BC during early morning sampling sessions in neighborhood areas
up to 300 m from the rail yard, as a function of distance from the rail yard. Concentration markers are located
at the midpoint of the distance range (e.g., 25 m for 0-50 m). Points for a given line are normalized by the
highest excess BC value for that data series.
•Evening-1
Evening-4
Evening-2
Evening-5
•Evening-3
Normalized
Excess BC
0.60
0.40
0.20
0.00
0 100 200 300
estimated distance from rail yard boundary (m)
Figure 3-22. Normalized downwind excess BC during evening sampling sessions in neighborhood areas up to
300 m from the rail yard, as a function of distance from the rail yard. Concentration markers are located at
the midpoint of the distance range (e.g., 25 m for 0-50 m). Points for a given line are normalized by the
highest excess BC value for that data series.
54
-------
As detailed in the Quality Assurance Project Plan (QAPP - Appendix A), 2-week QC checks and 6-
month performance evaluations were performed to meet DQOs. These checks are listed in section
2.4.4. Data completeness for each instrument is provided on a monthly basis in Table 3-5 and a
broad view of the data collected is provided in Figure 3-23. Despite the high completeness in data,
the carbon monoxide values were below detection limits (300 ppb) for 89% of the measuring period
(Figure 3-23). It should be noted that CO instrument detection limit is far below the regulatory limit
for CO (9 ppm at an 8 hr average or 35 ppm at a one hr average) and is commonly used for ambient
monitoring. At the time of the study, a CO instrument with a lower detection limit was not available
as a substitute and the very sensitive quantum cascade laser CO instrument used on the mobile
platform was isolated to the mobile sampling vehicle. Given the very low concentrations observed
routinely below the detection limit, CO is not included in any of the analyses to follow. The air
conditioning unit at the site began to fail in late May to early June of 2011, which caused a period of
overall data loss and led to significant instrument performance issues for the particulate matter
instruments. While monitoring utilizing all working instruments continued until October of 2011,
this analysis is currently restricted to the period of time when all measurements were being
collected simultaneously and the time span where operating conditions did not cause potential
concern for data. Therefore, the analyses in this report include only values up to May 5. Future
research involving these data may include data collected after that time, however the data will need
to be carefully reviewed to ensure instrument performance was acceptable. In addition,
NO/NO2/NOX data were invalidated from March 1 to March 13 because the QC check on March 13
did not meet the requirements set forth in the QAPP. Finally, it should be noted that PM2.s
measurements were collected but will not be presented, aside from brief discussion in the mobile
and stationary intercomparison section. During the time of available data, the instrument
performance met QA requirements, but the higher time resolution data appears to require post-
averaging in order to resolve the signal to noise, related to the tape-based method. The PM data
will require further analysis to determine the appropriate time-averaging interval.
An independent Technical System Audit and site evaluation were performed on February 8, 2011 by
Basim Dihu of USEPA Region 5.
55
-------
Table 3-5. Stationary monitoring data completeness by month
Year
2010
2011
2011
2011
2011
2011
2011
2011
2011
2011
2011
Month
November
January
February
March
April
May
June
July
August
September
October
CO
99.6%
99.9%
99.5%
99.8%
99.4%
87.3%
26.9%
95.4%
99.2%
92.4%
94.2%
NO
28.9%
100%
99.7%
99.1%
98.9%
93.5%
85.2%
50.6%
87.7%
92.3%
94.7%
NO2
28.9%
100%
99.7%
99.1%
98.9%
70.8%
49.3%
50.6%
87.7%
92.3%
94.7%
NOX
28.9%
100%
99.7%
99.1%
98.9%
93.5%
58.2%
50.6%
87.7%
92.3%
94.7%
S02
99.7%
99.9%
99.8%
99.8%
99.2%
93.6%
58.4%
94.7%
89.6%
92.4%
94.6%
wsa
29%
100%
100%
100%
100%
100%
96.5%
96.3%
83.2%
47.1%
95.6%
WDa
29%
100%
100%
100%
100%
100%
96.5%
96.4%
83.2%
47.1%
95.6%
BC
98.7%
97.8%
75.9%
99.3%
89.2%
70.4%
0%
0%
0%
0%
0%
PM2.5
90.2%
88.4%
66.2%
80.9%
100%
89%
54.7%
0%
0%
0%
0%
Gaps in MetOne meteorology (WS and WD) were filled using the meteorology from the PM2.5 E-BAM.
56
-------
Summary of concentration data
DEC
Jan
Feb
May
40 -
30 -
20 -
1D -
0 -
en
-i—•
o
'U
0.
l i i i i
50 100 150 200 250
100
80 -
60 -
40 -
20 -
o -
_
I I I
0.4 0.6 O.S
date
value
Figure 3-23. Time series and histogram summarizing stationary monitoring data collected during the time
period isolated for analysis (November 2010 - May 2011).
57
-------
Table 3-6. Summary statistics for 5-minute pollutant data (NOx and SO2 in ppb, BC in ng m"3)
Percentiles
Case
All data
Wind from SE
(angles: 105 -
215)
Wind from
SW
(angles:215 -
266)
Wind from N
(angles: 300 -
60)
Pollutant
N02
NO
NOX
S02
Black carbon
N02
NO
NOX
SO 2
Black carbon
N02
NO
NOX
SO 2
Black carbon
N02
NO
NOX
S02
Black carbon
N
38253
38253
38253
50085
47067
8274
8274
8274
8310
7652
7156
7156
7156
7176
6791
13212
13212
13212
13248
12235
Mean Standard
deviation
20.9
16.8
37.6
2.8
635.5
24.6
24.4
48.9
4.8
819.1
27.6
30.8
58.5
3.5
815.9
16.8
8.5
25.3
1.3
333.6
12.7
25.5
35.6
3.3
690.5
12.7
32.2
41.5
3.8
737.0
15.6
34.7
47.8
3.7
726.6
9.7
11.2
19.1
1.0
327.5
Lower
95th Cl
20.7
16.5
37.3
2.7
629.3
24.3
23.7
48.0
4.7
802.6
27.3
30.0
57.4
3.4
798.6
16.6
8.3
24.9
1.3
327.8
Upper
95th Cl
21.0
17.0
38.0
2.8
641.7
24.9
25.1
49.8
4.9
835.6
28.0
31.6
59.6
3.6
833.1
16.9
8.7
25.6
1.3
339.4
25th
11.2
3.0
15.7
1.0
235.4
15.2
3.7
21.2
2.3
378.0
15.4
6.4
23.7
1.3
359.0
9.6
2.4
13.2
0.8
143.3
50th
18.5
7.7
26.8
1.6
433.0
23.2
12.0
36.8
3.6
618.0
26.8
18.8
46.9
2.1
630.0
14.7
5.3
20.5
1.1
243.0
75th
28.6
18.6
46.9
3.2
796.0
32.7
32.1
63.8
5.8
1011.0
38.2
43.4
80.4
4.3
1031.5
22.5
10.1
32.0
1.5
419.0
90th
38.4
43.8
80.2
6.2
1330.0
42.5
63.8
104.2
9.9
1580.0
48.0
74.8
119.4
8.2
1610.0
31.1
18.8
48.2
2.4
662.0
95th
45.0
66.2
108.5
9.2
1811.0
47.4
92.1
134.2
12.5
2079.5
54.0
101.1
151.4
10.9
2092.0
35.1
27.8
59.8
3.1
858.3
Summary statistics for different wind sectors show that values are generally low for all pollutants
(Table 3-6). Confidence intervals for NOX compounds do not overlap for any of the sub-sectors, with
the highest values coming from the body of the rail yard (SW) and the lowest levels coming from
upwind. SO2 values from Midway airport (SE) might be significantly different from those of the rail
yard and upwind, but the absolute differences are small. Black Carbon values are over double, on
average, during winds from the South in comparison to wind from the North. CO was not included
in the table given that the significant fraction of time measurements was below the detection limit.
3.2.1.2. Data collection in context of local meteorology
Local wind trends can significantly influence local air quality trends for monitoring stations in close
proximity to emissions. The stationary monitoring site located to the northwest of the Cicero rail
yard collected wind data at a 5-minute resolution throughout the course of the study. These data
are compared against a similar timeframe at the Midway airport, located several miles south from
the stationary monitoring site. As shown in Figure 3-24, both monitoring stations show a significant
portion of the sampling year with winds from the West to South sector. Greater maximum and
mean wind speeds were evident at Midway in comparison to the stationary site. In addition, the
wind field at the Cicero site appears to have a greater SW-NE prevailing direction in comparison to
58
-------
the more broadly distributed flow at the Midway site. While no air flow obstructions were in the
immediate proximity of the stationary site at Cicero, it is possible that nearby buildings may have
influenced general local airflow.
Wind rose ol 5-nwiute average Cicero sue 2010-12-23 lo 2011-10-01
.
'
D-2 2-*
frequency ol cowtt by vrfndOrecttonTO
Wind rose of 2-mmute roiling averages Midway 2010-12-23 to 2011-05-05
Frequency 0* cwmts try wind flireclion (S)
Figure 3-24. Wind trend comparison for the Cicero monitoring site compared with the Midway airport
meteorological station.
Wind trends at the Cicero site were also analyzed as a function of time of day (Figure 3-25). Diurnal
patterns at the Cicero site show a greater frequency of winds from the east, from Lake Michigan,
during the day than at night. In addition, lower wind speeds (0-2 m/s) were more frequent during
the night time hours.
59
-------
Diurnal Wind rose of 5-minute averages: Cicero site 2010-12-23 to 2011-10-01
Frequency of counts by wind direction (S)
Figure 3-25. Cicero stationary monitoring site wind trends during the day (left figure) and night (right figure).
3.2.2. Assessment of rail yard impact
3.2.2.1. Impacts of meteorology and time of day on air quality measurements
Measured concentrations at the stationary site were evaluated as a function of wind direction, with
both sets of variables measured at five minute intervals. Shown in Figure 3-26, concentration roses
were created by binning the measured values according to wind direction, then magnitude (by 0-
25th, 25-50th, 50-75th, 75-95th, and 95-100* percentiles) and displaying in a polar plot with the length
of the ray relating to the frequency of that magnitude.
60
-------
POMKMIROSC
(c)
PoBufcm Rase B&cfc
frequency d ra*4B t>i iw* dUcction (N
(b)
dcowMby wwJ dir»tlwn t
(d)
Figure 3-26. Concentration roses for sulfur dioxide (ppb) (a), nitrogen oxide (ppb) (b), black carbon (ng m" ) (c),
and nitrogen dioxide (ppb) (d). The extent of a given ray indicates the fraction of data associated with a certain
wind direction and the rays are colored by concentration range measured from that particular wind sector (color
bins are: 0-25*, 25-5(f, 50-75*1, 75-95*, 95-100*percentiles).
Observing the relationship of concentrations with wind direction, SO2, NO, NO2 and black carbon all
appear to show strong directionality from the south and southwest with moderate levels also occurring
from the east. While the rail yard is located immediately south of the site, with a rail line oriented West-
East passing south of the site, additional industrial sources are present at a further distance to the south
(e.g., airport, power plant).
Wind directional trends in measured pollutant concentrations were also assessed as a function of time
of day (Figure 3-27). All pollutants show similar patterns for directionality and magnitude between day
and night, and show higher values to the southwest, south and east. Despite the lower wind speed
during the evening hours (Figure 3-25), there does not appear to be a significant increase in the general
measured concentrations during evening hours. The lower advection by wind may be offset by also
generally lower emissions activity during the evening hours.
61
-------
Diurnal pollution Rose: Sulfur Dioxide
Diurnal pollution Rose: Nitrogen Oxide
daylight
nighttimE
9.2132-40.225
3.202B-9.2132
1.6086-3.2Q2E
0.9934-1 .SOE6
O.OOSS-0.9934
SOj
Frequency of counts by wind direction (%)
Diurnal pollution Rose: Nitrogen Dioxide
nighttime
45.098-119.4
28.545-45.09S
18.625-28.545
11.402-18.625
0-11.402
Frequency of counts by wind direction (
Diurnal pollution Rose: Black Carbon
agylight
righttirre
t •*.<
.•'
w
f*
64.823-38S.2S
18.455-64.823
7.7656-1 S.455
3.0899-7.7656
| 0-3.0899
HO
1811-23821
796-1811
433-796
235.36-433
^tO-235.36
Blacfc.Carbon
Frequency of counts by wind direction (%
Frequency of counts by wind direction (%}
Figure 3-27. Diurnal concentration roses for sulfur dioxide (ppb), nitrogen oxide (ppb), nitrogen dioxide (ppb), and black carbon (ng m ). The extent of a given
ray indicates the fraction of data associated with a certain wind direction and the rays are colored by concentration range measured from that particular wind
sector (color bins are: 0-25*1, 25-50*, 50-75*, 75-95*, 95-100*percentiles).
62
-------
Daily patterns for pollutants were normalized by calculating the average value for each hour of the day
and dividing values by the mean of the set to normalize them. Bootstrapped 95% confidence intervals
were calculated for each hour. Shown in Figure 3-28, NO, NO2 and BC follow similar trends with peaks
in the morning and afternoon, indicating they are driven by commuter traffic. SO2 values peak at noon,
with the greatest mid-day means occurring from the east, southeast, south and southwest. This
indicates that Midway Airport and the Crawford power station are the potential major contributors to
mid-day SO2.
so.
NO,
NO
Black. Carbon
1.6 -
1.4 -
1,2 -
TO
O
c
1,0 -
0.8
Figure 3-28. Normalized diurnal time series of measured concentrations at the stationary monitoring site, for all
data collected during November, 2010 - May, 2011.
63
-------
so.
NO,
NO
Black.Carbon
20 -
1.5 -
1.0 -1
0.5 -
Figure 3-29. Normalized diurnal time series of measured concentrations at the stationary monitoring site, for
weekday (left) and weekend periods (right) collected during November, 2010 - May, 2011.
Using the same method pollutants were separated into weekdays and weekends and hourly means
were plotted, shown in Figure 3-29. SO2 shows similar patterns on weekdays and weekends. Other
pollutants show overall lower values and depressed peaks on weekends. The change in pattern for NO,
NO2 and BC is consistent with the differences in commuter traffic on weekends.
64
-------
25 -
2.0 -
0>
-c>
1.0 -
0.5 -
S02
NO.
NO
BlacK,Carbon
Figure 3-30. Normalized diurnal time series of measured concentrations at the stationary monitoring site, for
periods of wind from the North (left) and wind from the South (right) collected during November, 2010- May,
2011.
By the same method but binning concentrations by the northern sector (less than 90 degrees or
greater than 270 degrees) and southern sector (greater than 90 degrees and less than 270 degrees)
concentrations show similar patterns with significantly greater values when winds are from the
south (Figure 3-30).
65
-------
The results for the black carbon analysis are shown in Figure 3-31. Black carbon is the directly
emitted product of fossil fuel combustion, especially from diesel emissions which relates to the truck
and locomotive operations at the rail yard. The concentration field shows that the higher black
carbon concentrations are confined mostly to times when air parcels pass over land south of the
monitor's location with notably higher concentrations occurring when winds are from the southwest
and southeast. The highest concentrations (greater than 1 u.g/m3) occur when the winds pass over
areas due south and adjacent to the rail yard but also from the area around the Crawford power
plant and another large rail yard which is due south of Crawford on the other side of the Chicago
Sanitary and Ship Canal. The area around the Cicero stationary monitor is heavily laden with fossil
fuel burning industries. The NTA does an adequate job of isolating the majority of peak black carbon
emissions to the southern half of the grid with isolating nearby local sources of fossil fuel
combustion as possibly being the largest contributors to black carbon at the Cicero monitoring site.
The NTA shows lower level expected black carbon concentrations north of the monitor site including
areas along major roadways such as 1-290. These expected concentrations are much more uniform
over a large area suggesting more consistent emissions from dense urban mobile sources usually
present in a large metropolitan area. The black carbon pollution rose in Figure 3-26 shows a
consistent pattern with the highest concentrations south of the monitor and lower concentrations
to the north. Such a gradient indicates that sources south of the site contribute more to the higher
black carbon observations measured by the monitor than general urban mobile source emissions
consistent across large metropolitan areas.
66
-------
Figure 3-31. Black carbon expected concentration field from NTA in ng/m
The SO2 expected concentrations show a pattern similar to black carbon in that the majority of SO2
emissions appear to be coming from the southern half of the domain (Figure 3-32). The higher
concentrations are isolated along heavily traveled transportation corridors and industrial zones within
the area including Interstate 94 (Dan Ryan Expressway), south of the monitor with air parcels passing
over Midway Airport and another large rail yard close to Midway, and the Interstate 55 corridor along
the McCook industrial area by the Chicago Sanitary and Ship Canal. The SO2 expected concentration
field also shows "coning" to the south of the monitor as concentrations increase toward Midway Airport
and suggests that persistent winds from the southerly direction could be representing emissions not
only attributable to Midway but also nearby SO2 emissions from the other local sources including the
Cicero rail yard. The commingling of emissions from local sources in close proximity to one another
makes it difficult to determine the extent of the Cicero rail yard impact on ambient SO2 concentrations
at the receptor. However, the amount of air traffic from Midway and the emissions associated with jet
aircraft cannot be ignored as a major source of SO2 for the area.
67
-------
Figure 3-32. Sulfur dioxide expected concentration field from NTA.
The NO concentration field is more similar to that of black carbon suggesting that both NO and black
carbon are more likely to be coming from similar sources within the area (Figure 3-33). The Pearson
correlation coefficient shows that the black carbon and NO have the highest correlation to each other
out of the three possible pairings between black carbon, NO and SO2 (Table 3-7). While this correlation
could suggest that the black carbon and NO come from the same source, there are multiple mobile
sources within the area burning fossil fuels including various on-road diesel sources. The relatively low
correlations between the three pollutants suggest that area around the monitoring site is dominated by
a diverse variety of sources in close proximity to one another all contributing to ambient air quality by
varying amounts thereby making it difficult to discern a distinct signal from any one particular source.
The difference in the spatial variability in NO concentrations versus SO2 concentrations helps to validate
that aircraft emissions from Midway Airport are contributing substantially to the area's SO2
concentrations. The high NO concentrations in the vicinity of the Crawford power plant are also
consistent with emissions expectations for a source of its size. With the Crawford power plant
scheduled to be completely shut down by the end of 2014, the majority of black carbon, SO2 and NO
contributions to local air quality will further shift to transportation-oriented sources.
68
-------
Figure 3-33. Nitric oxide expected concentration field from NTA.
69
-------
Table 3-7. Pearson Correlation Coefficients (R) for Pollutant Pairs
Sulfur Dioxide
Nitric Oxide
Nitric Oxide
0.19
Black Carbon
0.31
0.52
70
-------
Elevated local air pollution levels within several hundred meters of a major transit source - also referred
to as near-source air pollution - have been observed in numerous field studies measuring air quality
alongside major roadways and motivates ongoing research to understand local air quality trends near
other transit modes. Rail yards are important nodes in the rail network, serving to route trains and
transfer containers from one mode of transportation to another. Emissions activities within typical
intermodal rail yards - such as the Cicero rail yard studied in this report - include commercial trucks
arriving and departing with containers, within-yard hostler trucks, switcher locomotives moving
containers within the yard, freight and/or commuter trains transiting through the yard, and diesel-
powered cranes. These emissions are heterogeneously distributed over a discrete large area and have
both temporal and spatial variability.
Rail yards are commonly located in densely populated areas, as the customer and industry base of
populated areas demands efficient transportation of goods. The Cicero rail yard is a good example of
this situation, where dense residential communities are built right next to the rail yard boundary. In
these environments, it is important to understand whether these communities are at a greater risk for
air pollution due to local emissions. This research study evaluated local air pollution trends both
spatially and temporally, utilizing a combined strategy of short-term mobile and longer-term stationary
monitoring.
Mobile monitoring results indicate that the intermodal rail yard area - which for this data set analysis
includes freight and commuter trains, trucks, cranes, and limited traffic on boundary public roads to the
North and South of the yard - is associated with elevated black carbon concentrations in downwind
residential neighborhoods up to and possibly exceeding 300 m in downwind distance from the rail yard
boundary. The attribution of this impact to the rail yard environment is most clear during the early
morning and evening time periods, when a clear difference in elevated BC exists between cases of winds
from the North versus South. During the mid-day, residential neighborhoods experience elevated BC
under multiple wind conditions, which suggests that other local sources of BC (e.g., street traffic) also
contribute to elevated concentrations in these neighborhood areas. Other pollutants measured on the
mobile platform - ultrafine particles, carbon monoxide, and particulate matter-either did not show
statistically significant impact or had inconsistent trends, likely explained by the pollutant emissions and
also affected by monitoring instrumentation limitations. Regarding pollutant emissions, abundant
directly emitted pollutants associated with diesel emissions are anticipated to be the most sensitive
tracers of local air pollution impact. As observed when sampling in stationary mode, BC spiked 45-fold
when experiencing a near-field emissions plume; meanwhile, carbon monoxide and PM2.s only changed
by a factor of 2 and UFPs by a factor of 11. This difference is likely due to variations in emission factors,
whether the pollutant has significant background levels (e.g., PM2.s anticipated to have a significant
regional component), as well as the sensitivity of the measurement instrumentation. Ultrafine particles
(particles smaller than 100 nm) often track closely with BC signals for highway environments; however,
emissions and field studies indicate that increased ultrafine particle emissions are related to higher
71
-------
driving speeds. The results of this study indicate that BC and UFPs may not track as closely in the rail
yard environment, but more research is needed to further understand this relationship.
Stationary monitoring results also indicate a clear association of air pollution levels with wind direction,
finding air pollution levels for multiple pollutants (BC, NO2, NO, SO2) elevated with winds from the
south. Nonparametric trajectory analysis - an inverse modeling research tool that utilizes high time-
resolution air monitoring and wind data - indicates that important source areas affecting stationary
monitoring data include areas to the Southwest (including the rail yard area), due South (including
Midway airport), and the Southeast (including a power plant). Diurnal analyses of stationary monitoring
results, isolated in weekday/weekend timeframes and North / South wind timeframes, indicate higher
overall pollution levels on weekdays and with winds from the south.
The study results overall indicate that residential areas in close proximity to the Cicero rail yard generally
experience higher overall air pollution levels with winds from the South, which may be related to
multiple significant emission sources including the rail yard environment. In addition, under southerly
wind conditions, mobile monitoring data suggests an indicator of diesel emissions (black carbon)
increased 30-104% over the urban background during early morning and evening periods and is more
directly associated with emissions activity associated with the rail yard area. Uncertainty remains
regarding source attribution, both within the rail yard and considering potential traffic on boundary
roads, which may require modeling or controlled field experiments for further characterization. These
results support the notion that local concentrated areas of higher diesel emissions activity adversely
impact local-scale air quality and mitigation efforts may reduce local exposure to air pollution.
5.
This research was conducted under EPA's Regional Applied Research Effort (RARE) program, which
encourages collaborative research between the EPA Regions and the Office of Research and
Development. Field measurement support for the mobile monitoring campaign was provided by
ARCADIS through contract EP-C-09-027. BNSF staff provided in-kind support to this project through
sharing rail yard activity data for the Cicero rail yard, providing logistical support for the mobile
monitoring study, and providing insight throughout the project. EPA QA personnel in ORD and Region 5
supported this study through reviewing Quality Assurance Project Plans and report drafts. EPA
researchers Ram Vedantham and Rich Baldauf are also appreciated for providing technical peer reviews
of a draft version of the report.
72
-------
6,
Baldauf, R., E. Thoma, V. Isakov, T. Long, J. Weinstein, I. Gilmour, S. Cho, A. Khlystov, F. Chen, J. Kinsey,
M. Hays, R. Seila, R. Snow, R. Shores, D. Olson, B. Gullett, S. Kimbrough, N. Watkins, P. Rowley, J.
Bang, and D. Costa. 2008a. Traffic and meteorological impacts on near road air quality: Summary
of methods and trends from the Raleigh Near Road Study. Journal of the Air & Waste
Management Association 58:865-878.
Baldauf, R., E. Thoma, A. Khlystov, V. Isakov, G. Bowker, T. Long, and R. Snow. 2008b. Impacts of noise
barriers on near-road air quality. Atmospheric Environment 42:7502-7507.
Barzyk, T. M., B. J. George, A. F. Vette, R. W. Williams, C. W. Croghan, and C. D. Stevens. 2009.
Development of a distance-to-roadway proximity metric to compare near-road pollutant levels
to a central site monitor. Atmospheric Environment 43:787-797.
Cahill, T. A., T. M. Cahill, D. E. Barnes, N. J. Spada, and R. Miller. 2011. Inorganic and organic aerosols
downwind of California's Roseville railyard. Aerosol Science and Technology 45:1049-1059.
Campbell, D., and Fujita, E. M. 2009. Roseville rail yard air monitoring project (RRAMP). Final report
summary of data QA and trend analysis., Desert Research Institute, Reno, NV.
Chicago Metropolitan Agency for Planning. 2011. Chicago Intermodal Facility Lift Counts and Regional
TED Estimate (Updated December 2011).
Federal Highway Association. 2010. Freight and Air Quality Handbook. Prepared by the U.S. Department
of Transportation, Federal Highway Administration, Washington, DC.
Hagler, G. S. W., R. W. Baldauf, E. D. Thoma, T. R. Long, R. F. Snow, J. S. Kinsey, L. Oudejans, and B. K.
Gullett. 2009. Ultrafine particles near a major roadway in Raleigh, North Carolina: Downwind
attenuation and correlation with traffic-related pollutants. Atmospheric Environment 43:1229-
1234.
Hagler, G. S. W., M.-Y. Lin, A. Khlystov, R. W. Baldauf, V. Isakov, J. Faircloth, and L. E. Jackson. 2012. Field
investigation of roadside vegetative and structural barrier impact on near-road ultrafine particle
concentrations under a variety of wind conditions. Science of the Total Environment 419:7-15.
Hagler, G. S. W., E. D. Thoma, and R. W. Baldauf. 2010. High-resolution mobile monitoring of carbon
monoxide and ultrafine particle concentrations in a near-road environment. Journal of the Air &
Waste Management Association 60:328-336.
Hagler, G. S. W., T. L. B. Yelverton, R. Vedantham, A. D. A. Hansen, and J. R. Turner. 2011. Post-
processing method to reduce noise while preserving high time resolution in Aethalometer real-
time black carbon data. Aerosol and Air Quality Research 11:539-546.
Hansen, A. D. A., H. Rosen, and T. Novakov. 1984. The Aethalometer - an instrument for the real-time
measurement of optical absorption by aerosol particles. Science of the Total Environment
36:191-196.
HEI Panel on the Health Effects of Traffic-Related Air Pollution. 2010. Traffic-related air pollution: A
critical review of the literature on emissions, exposure, and health effects. . Health Effects
Institute. Boston, Mass.
Henry, R. C., A. Vette, G. Norris, R. Vedantham, S. Kimbrough, and R. C. Shores. 2011. Separating the air
quality impact of a major highway and nearby sources by nonparametric trajectory analysis.
Environmental Science & Technology 45:10471-10476.
Hitchins, J., L. Morawska, R. Wolff, and D. Gilbert. 2000. Concentrations of submicrometre particles from
vehicle emissions near a major road. Atmospheric Environment 34:51-59.
Hu, S., S. Fruin, K. Kozawa, S. Mara, S. E. Paulson, and A. M. Winer. 2009. A wide area of air pollutant
impact downwind of a freeway during pre-sunrise hours. Atmospheric Environment 43:2541-
2549.
73
-------
Karner, A. A., D. S. Eisinger, and D. A. Niemeier. 2010. Near-roadway air quality: synthesizing the findings
from real-world data. Environmental Science & Technology 44:5334-5344.
R Development Core Team. 2008. R: A language and environment for statistical computing. Vienna,
Austria.
Turner, J. R., Yadav, V., and Feinberg, S.N. 2009. Data analysis and dispersion modeling of the Midwest
rail study (Phase I) -final
report, http://www.ladco.org/reports/general/new docs/WUSTL MidwestRailStudy FinalRepor
t.pdf.
Welch, B. L. 1947. The generalisation of student's problems when several different population variances
are involved. Biometrika 34:28-35.
Yamartino, R. J. 1984. A comparison of several single-pass estimators of the standard deviation of wind
direction. Journal of Climate and Applied Meteorology 23:1362-1366.
Zhu, Y. F., W. C. Hinds, S. Kim, and C. Sioutas. 2002. Concentration and size distribution of ultrafine
particles near a major highway. Journal of the Air & Waste Management Association 52:1032-
1042.
74
-------
Cicero Rail Yard Study (CIRYS)
Appendix A: Quality Assurance Project Plans
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Pagel of 28
v°/EPA
Cicero Rail Yard Study (CIRYS) - High-Resolution Mobile Monitoring of Near-
Rail Yard Air Pollution
Quality Assurance Project Plan - Revision 2
Category III / Measurement Project
NRMRL/APPCD/ECPB
EPA Technical Lead: Gayle Hagler
13 January 2011
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 2 of 28
Cicero Rail Yard Study (CIRYS) - High-Resolution Mobile Monitoring of Near-
Rail Yard Air Pollution
Quality Assurance Project Plan
Table of Contents
1. Approvals 3
2. Distribution List 4
3. Project Description and Objectives 5
3.1 Background 5
3.2 Project Purpose and Objectives 8
4. Organization and Responsibilities 8
4.1 Project Personnel 8
4.2 Project Schedule 9
5. Scientific Approach 9
5.1 Sampling Design 9
5.2 Process Measurements 15
5.3 General Approach and Test Conditions for Each Experimental Phase 15
6. Sampling Procedures 15
6.1 Site-Specific Considerations 15
6.2 Sampling Equipment and Procedures 16
6.3 Quality Control in Sample Analysis 19
6.4 Sample Preservation 19
6.5 Sample Numbering 19
6.6 Sample Chain-of-Custody 19
7. Measurement Procedures 20
7.1 GMAP monitoring 20
7.2 Calibration Procedures 21
8. Quality Metrics 22
8.1 QC Checks 22
8.2 QA Objectives and Acceptance Criteria 23
9. Data Analysis, Interpretation, and Management 25
9.1 Data Reporting 25
9.2 Data Validation 25
9.3 Data Analysis 25
9.4 Data Storage Requirements 27
10. Reporting 27
10.1 Deliverables 27
10.2 Expected Final Products 27
11. References 27
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 3 of 28
1. Approvals
Gayle Hagler Date
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Project Leader/ARCADIS Work Assignment Manager
Eben Thoma Date
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Project Scientist/Alternate ARCADIS Work Assignment Manager
Robert Wright Date
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Quality Assurance Representative
David Proffitt Date
ARCADIS
Work Assignment Leader
Libby Nessley Date
ARCADIS
Quality Assurance Representative
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 4 of 28
2. Distribution List
Bob Wright - EPA ORD
GayleHagler-EPAPRD
Eben Thoma - EPA ORD
Richard Shores - EPA ORD
Monica Paguia - EPA Region 5
Loretta Lehrman - EPA Region 5
David Proffitt-ARCADIS
Richard Snow - ARCADIS
Michal Derlicki - ARCADIS
Libby Nessley - ARCADIS
Paul Nowicki - BNSF
David Seep-BNSF
Michael Stanfill - BNSF
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 5 of 28
3. Project Description and Objectives
3.1 Background
Air quality research has progressed over recent years from focusing on primarily regional level
air pollution (1 Os of kilometers), monitored using a network of ambient monitoring sites that are
removed from any major sources, to an emerging parallel focus on local air pollution (10s of
meters). In the transportation sector, recent field studies have shown that air pollution levels
can be significantly higher than background levels when immediately downwind of a highway
(e.g., Zhu et al., 2002), airport (e.g, Westerdahl et al., 2008), or rail yard (e.g., Chang, 2007).
While a large number of field studies have taken place to assess local air pollution effects from
highway-related emissions, air monitoring data are sparse for many other sources, including
distribution centers, airports, rail yards, refineries, powerplants, ports, etc. This study is seeking
to begin to fill in the knowledge gaps by focusing on studying local impact due to rail yard
emissions.
Near-road research, which is anticipated to have similarities for near-rail yard research, has
quantified elevated concentrations on the downwind side of the source attenuating with
increasing distance and eventually reaching background levels (e.g., Baldauf et al., 2008a).
Near-road field studies have determined that the degree of air pollution concentration elevation
over background concentrations and the rate of attenuation with distance from a road depends
on the pollutant type, meteorology, and the surrounding terrain. For example, a near-road
study in North Carolina revealed that certain air pollutants, such as black carbon, nitrogen
oxide, and ultrafine particles (particles with diameters smaller than 100 nanometers), have a
stronger response to traffic emissions in comparison to other air pollutants, such as fine
particulate matter (particles with diameters smaller than 2.5 micrometers) or nitrogen dioxide
(Hagler et al., 2009, Thoma et al., 2008). Another study in California demonstrated the
significant effect local meteorology can have on the downwind dispersion of traffic-related air
pollution, with an atmospheric inversion leading to elevated concentrations up to a mile and a
half in distance from a major highway (Hu et al., 2009). Finally, near-source obstacles, such as
buildings and walls, have been shown to alter the concentrations of near-road air pollution
(Baldauf et al., 2008b, Heist et al., 2009). While these near-road studies have identified several
important factors affecting emissions dispersion, it is an open question as to what degree the
pollutant type, local meteorology, and surrounding terrain determine concentration gradients for
other source types, including rail yards.
This study seeks to characterize local impacts from an intermodal rail yard in Cicero, IL,
including magnitude of impact for specific pollutants, assessing the spatial variability of
concentrations, and downwind extent of elevated concentrations over the upwind background
levels. This research study was originated by an EPA Region 5 RARE proposal, which
hypothesized that rail yard emissions may locally elevate air pollutant levels in their region and
contribute to regional PM25 nonattainment status. Phase I of the study focused on the CSX
Rougemere Rail Yard in Dearborn, Michigan in a multi-component research effort that included
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 6 of 28
a rail yard emissions inventory, dispersion modeling, and a several month monitoring campaign
measuring carbonaceous particulate matter (black carbon, elemental carbon, and organic
carbon) at two sites immediately adjacent to the rail yard and additional site located within a
nearby residential area and representative of urban background concentrations. The results
from the Phase I study have been published in a report (Turner et al., 2009). The emissions
inventory component of this study identified a number of different emission types within the rail
yard, as shown in Table 3-1 and Figure 3-1 (Heiken, 2009) and demonstrated a reduction in
PM2 5 emissions due to a switch to lower sulfur fuel in 2008. Analysis of ambient monitoring
data collected upwind and downwind of the Dearborn rail yard did not successfully isolate the
rail yard signal due to other large emitting sources immediately adjacent to the rail yard. The
Phase I experience lead to the Phase II measurement campaign being located at another study
site (Cicero, IL) that met site criteria including few confounding sources (Section 5).
Table 3-1. Emission Sources at the CSX Rougemere Rail Yard (Heiken, 2009)
Locomotives
• Through Locomotives
• Arrival/Departure Locomotives
• Additional Idling from
Arrival/Departure Locomotives
• Switcher Locomotive Operation
• Switcher Refueling Idling
Non-Loco motives
• Worker Vehicle Exhaust
Worker Vehicle Evaporative
• HDDT Delivery
• HDDT Delivery idle
• Facility Truck Facility Truck Idle
Space Heating
• Water Heating
• LPG - Welders/Cutters
Diesel - Specialty Vehicle Carts
• Diesel - Rubber Tire Loaders
• Diesel - Forklifts
Diesel - Other General Industrial
Equipment
• Diesel - Srtowblowers
• Aerosol Paints
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 7 of 28
l.oo
0.80
in
O 0.60
0.40
n2007annual (t.Stpy)
• 2008 annual (1.0 tpy)
0.20
0.00
Through Arrival/Departure Additional Idling Switcher Switcher Refueling Non-Locomotive
Locomotives Locomotives fromArr/Dep Locomotive Idling
Locomotives Operation
Figure 3-1. PM25 emissions by source type at the CSX Rougemere Rail Yard in Dearborn,
Michigan (Heiken, 2009).
Rail yard emissions are unevenly distributed over the source area, with certain zones with high
activity (e.g., container moving by cranes, through locomotives) and other zones relatively
inactive (e.g., container storage). This contrasts to the well-studied highway source, which is a
generally homogenous spatially distributed line source that varies temporally. Rail yard
emissions vary both temporally and spatially, although limited by known boundaries. This study
seeks to add to the body of knowledge on rail yard impacts on local air pollutant levels in the
Region 5 territory, and is focused on the BNSF Cicero Rail Yard in Cicero, IL, as a fairly
representative example of an intermodal rail yard. This study will use a mobile monitoring
approach, combined with local meteorology information, to collect data on air pollutant
concentrations upwind and downwind of the rail yard. This method utilizes the flexibility of
mobile monitoring to collect data surrounding the rail yard and at extended distances downwind
of the rail yard, allowing upwind/downwind concentrations to be compared and the extent of
downwind influence to be detected. Recognizing the diurnal variability in the atmospheric
mixing height, local wind conditions, and rail yard emissions activity, mobile deployments will be
conducted over several different time frames to observe the variability of near-rail yard air
pollutant concentrations and spatial extent of any excess air pollution detected over the upwind
background levels. The different time frames will also cover periods anticipated to range from
low to moderate side road traffic, thus will allow the influence of this other local source on data
to be detected.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 8 of 28
3.2 Project Purpose and Objectives
The purpose of this study is to measure the spatial patterns of rail yard-related gas- and
particle-phase pollutants at a rail yard site within Region 5.
The specific objectives of this research study are:
• Measure the spatial extent of local air pollution elevated over the background,
downwind of a major rail yard in Region 5.
• Measure the spatial and temporal variability of near-rail yard air pollution, under
different meteorology conditions and source emission characteristics.
4. Organization and Responsibilities
4.1 Project Personnel
This QAPP addresses the measurement of spatial patterns of ambient air pollution nearby
major rail yard using air monitoring instruments onboard multiple mobile and fixed sampling
units. The field measurements will be performed by ARCADIS personnel, with technical
guidance from EPA/ORD/NRMRL personnel. Table 4-1 lists the personnel responsible for the
oversight and QA review of this project. Additional team members that will be involved in this
study are listed in Table 4-2.
Table 4-1. Key Points of Contact
Name
Gayle Hagler
Eben Thoma
Bob Wright
David Proffitt
Libby Nessy
Organization
Affiliation
EPA/NRMRL
EPA/NRMRL
EPA/ NRMRL
ARCADIS
ARCADIS
Title
Technical
Leader/Project
Leader
Project
Scientist
QA Manager
Work
Assignment
Leader
ARCADIS QA
Manager
Responsibilities
Technical leadership for
study, WAM for
ARCADIS research
support, data analysis
for fixed and backpack
monitoring
Technical support for
study, Alternate WAM
for ARCADIS research
support, data analysis
for mobile monitoring
Quality Assurance
through review of QAPP
and presentation of
results
Coordination of
ARCADIS research
support activities
ARCADIS QA activities
oversight
Contact Information
Phone:(919)541-2827
Email: hagler.gayle@epa.gov
Phone:(919)541-7969
Email: thoma.eben@epa.gov
Phone: (919)541-4502
Email: wright.bob@epa.gov
Phone: (91 9) 544^535
Email: david.proffitt@arcadis-
us.com
Phone:(919)541-2260
Email: lnessley@arcadis-us.com
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 9 of 28
Parik Deshmukh
Monica Paguia
Loretta Lehrman
ARCADIS
EPA Region 5
EPA Region 5
Field operator
Project team
member
Project team
member
Operation of air
monitoring vehicle,
instrumentation, and
data management.
Region 5 project
leadership.
Region 5 project
coordination.
Phone: 919-541-2980
Email:
Parikshit.Deshmukh@arcadis-
us.com
Phone:312-353-1166
Email: paguia.monica@epa.gov
Phone: 312-886-5482
Email: lehrman.loretta@epa.gov
Table 4-2. Additional Project Team Members
Name
Bill Mitchell
Bill Squier
James Faircloth
Jerry Faircloth
Michal Derlicki
Richard Snow
Chad McEvoy
Jaime Wagner
Organization
Affiliation
EPA/NRMRL
EPA/NRMRL
EPA/NRMRL
ARCADIS
ARCADIS
ARCADIS
EPA Region 5
EPA Region 5
Title
Project
Engineer
Project
Engineer
Project
Technician
Project
Technician
Project
Technician
Project
Scientist
Project team
member.
Project team
member.
Responsibilities
Data-logging set-up
support
Shop support for
instrumentation mounting.
GMAP vehicle
troubleshooting support
Field study support
Field study support
Field study support
Field study support
Field study
communications
Contact Information
Phone:(919)541-2515
Email: mitchell.bill@epa.gov
Phone: (919)541-2516
Email: squier.bill@epa.gov
Phone:(919)541-0157
Email: faircloth.james@epa.gov
Phone: (919)541-0314
Email: faircloth.jerry@epa.gov
Phone:(919)544-4535
Email: Michal. derlicki@arcadis-
us.com
Phone: (919)541-3135
Email: snow.richard@epa.gov
Phone:(312)886-6084
Email:
Mcevoy.chad@epa.gov
Phone:(312)886-9402
Email:
Wagner.jaime@epa.gov
4.2 Project Schedule
The target project schedule is as follows:
• September, 2010: Preparation activities for field campaign
• October-November, 2010: Field campaign at the Cicero Rail Yard.
• December, 2010-March, 2011: Preliminary data analysis, including data report.
• March, 2011-September, 2011: Ongoing data analysis and presentation of results.
5. Scientific Approach
5.1 Sampling Design
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 10 of 28
In order to meet project objectives (Section 3.2), the following field data will be collected:
1) High-resolution mobile monitoring campaign - EPA's Geospatial Monitoring of Air
Pollution (GMAP) vehicle will be deployed to map air pollutants surrounding the rail
yard boundary and in near-rail yard neighborhoods.
2) Local meteorology data will be collected using a portable meteorology station.
The field study will be conducted over a 1-month period in Cicero, Illinois. The length of the
field campaign was chosen to allow for repetitive fixed and mobile monitoring under a variety of
meteorology conditions.
5.1.1 Site Selection and Description
The monitoring sites are selected based on the following criteria, shown in Table 5-1.
Table 5-1. Site selection criteria
Criteria
Activity level of rail yard
Existence of historical monitoring data at the site
Ease of setting up a fixed sampling site and monitoring meteorology and air
quality for several months.*
Few other nearby sources
Capability to drive in close proximity to rail yard on multiple sides, particularly
along axis of prevailing wind.
Access to low traffic roads surrounding rail yard, to avoid biases from single
vehicle exhaust.
Characteristics of surrounding environment (residential, commercial, etc.)
Rank(H:high,
M:mid, L:low)
H
M
H
H
H
H
M
To support collaborative monitoring effort by Region 5 staff.
A number of monitoring sites were considered throughout the Chicago-area in Region 5,
including the following rail yards - Corwith, Proviso, Cicero, 59th Street, The Belt, Ashland,
and Elwood. The BNSF Cicero Rail Yard was selected as the optimal site based upon the
criteria laid out in Table 5-1 and is shown in Figure 5-1. Cicero is an intermodal rail yard,
with both locomotive and truck traffic inside the rail yard boundary. A common metric for
intermodal rail yard activity is in terms of shipping container lifts per day - Cicero has
approximately 1000-1200 lifts per day. In addition to emissions by diesel-powered cranes,
other on-site emissions include 5-9 hostler trucks, 500 in/out truck traffic, 8 daily intermodal
trains, and approximately 140 through trains (~120 passenger trains, 20 mixed freight
trains), and 4-5 switcher locomotives.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 11 of 28
Figure 5-1. Satellite image of the Cicero Rail Yard.
The area surrounding the Cicero Rail Yard is primarily residential neighborhoods, which
provides a dense network of low-traffic residential roads allowing for multiple transects to
be driven upwind and downwind of the rail yard. The roads surrounding the rail yard and
estimated Average Annual Daily Traffic (AADT) for 2009 by the Illinois DOT are shown in
Table 5-2 and Figure 5-2.
Table 5-2. Daily traffic counts on nearby roadways
Street
Cicero Ave. (South of rail line)
Cicero Ave. (North of rail line)
Cicero Ave. (Further south near I-55)
26tn North of yard
Ogden (east of yard)
Ogden (west of yard)
Ogden to Cicero turn off
Austin (west end of yard)
I-55 (~2 mile south)
Truck only
(Daily counts)
9,100
2,200
5,050
N/A
N/A
N/A
N/A
N/A
12,000
Total Traffic
(Daily counts)
43,200
34,300
41,300
9,900
22,400
18,500
11,200
13,700
149,100
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 12 of 28
no is DOTtg NAVTEQ 2005
Figure 5-2. Average Annual Daily Traffic (AADT) along roads surrounding the Cicero Rail Yard.
The surrounding roads generally have low to moderate level traffic (150,000 AADT is the
threshold usually considered as indicating a major highway) and the nearest major highway
is over a mile away. In the interest of minimizing potential biases due to the roadways with
moderate level traffic (e.g., Ogden Ave, Cicero Ave), the sampling will be conducted
outside of typical commuter hours. However, it should be noted that up to 20% of traffic on
some of the busier arterials can be due to truck traffic, which may not follow traditional
commute hours. In addition to avoiding hopefully the worst case time periods, another
preventative action to minimize bias due to local truck exhaust is the selection of specific
residential side roads for monitoring without significant truck traffic. Finally, the electric
vehicle will have a webcam recording the driver's view as well as real-time instruments that
can detect sudden spikes in concentrations that would indicate a local exhaust event. The
driving route is provided in more detail in section 5.1.3 and the monitoring actions to
mitigate local exhaust bias are detailed in section 6.1.
5.1.2 Sampling Schedule
The mobile monitoring is conducted using an electric vehicle outfitted with fast-response air
monitoring instruments. This vehicle has an on-board battery supply supporting driving and
powering the air monitoring instruments. The daily sampling duration is limited by power
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 13 of 28
availability, usually limited to approximately 2-3 hours of driving mode sampling, followed by
1-2 hours of stationary sampling. Several considerations are guiding the selected days of
the week and time of day for sampling:
• Rail yard activity: BNSF reports that the lowest rail yard activity days are Tuesday and
Wednesday, with Thursday, Friday, and Saturday having the highest traffic. During the
day, highest traffic occurs during 10-11:30 AM and 8-9 PM. Lowest traffic occurs from
midnight - 8 AM.
• Avoidance of commuter traffic: Commuter traffic windows (6:30-8:30 AM, 4:30-6:30
PM) are to be avoided in order to minimize the degree of local traffic influence.
• Meteorology: Atmospheric mixing height is lowest in the evening and pre-sunrise
hours, which can reduce the dispersion of emissions and increase ground-level
concentrations. During the day, the atmospheric mixing height increases as the sun
heats the ground surface and dispersion of pollutants increases.
The following three periods of time are selected for sampling activity, all outside of the main
commute periods:
A) 7 PM - 9:30 PM: Peak traffic at rail yard, mid-level atmospheric mixing height
B) 9 - 11:30 AM: Peak traffic at rail yard, high atmospheric mixing height
C) 4 AM - 6-30 AM: Low traffic at rail yard, low atmospheric mixing height
Sampling will occur over 6 days per week, and a tentative sampling schedule is laid out in
Table 5-3. The schedule rotates the deployments as A, B, C, repeat, with a goal of having
8 of each time schedule and a variety of wind conditions for each time window. The
rotation scheme is also considering the need fora vehicle recharge window between
deployments, providing a minimum of 10 hours between deployments. An example
sampling deployment timing is provided in Table 5-4.
Table 5-3 Tentative Sampling Schedule (blue = 7-9:30 PM [A], orange = 9-11:30 AM [B], gray = 4-6:30
AM [C])
OCTOBER
M
4
II
18
25
T
5
12
19
26
W
6
13
20
27
T
7
14
21
28
F
1
8
15
22
29
S
2
9
16
23
30
S
3
10
17
24
31
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 14 of 28
Table 5-4 Daily Sampling Schedule (e.g. for Schedule C)
Time
0330
0400
0630
0800
Action
GMAP vehicle instrument QC checks
Monitoring initiates - GMAP vehicle on route (4-6 repeats)
GMAP driving ends, stationary sampling initiates
GMAP stationary sampling ends, data downloaded
5.1.3 Driving Route
The electric car driving route outside of the rail yard boundaries is shown in Figure 5-3.
This driving route was designed to have multiple transects extending at least 200 m away
from the rail yard on each side of the yard. Several transects extend well over 300 m in
distance, which is the threshold distance at which near-road field studies typically see
elevated concentrations return to background levels. Data at distances downwind and data
collected upwind of the rail yard will be compared to determine the distance at which
concentrations downwind are similar to that upwind.
Figure 5-3. External rail yard mobile monitoring driving route (blue line) and location for
stationary monitoring (yellow marker).
The draft driving route is approximately 11 miles and is anticipated to take 25 minutes to
complete and allowing for the route to be repeated multiple times within the 2-3 hour driving
period. This route may also be extended, or alternate with, a section of road that is within
the rail yard boundaries, upon consultation with BNSF for access and vehicle safety. The
driving route will be tested during a pre-deployment site visit for feasibility and may be
altered, but will meet the goal having multiple transects on each side of the rail yard. The
selection of the stationary monitoring location, to occur at the end of the driving route, is
based upon Region 5's plan to initiate a stationary monitoring site there to collect data
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 15 of 28
continuously during the field campaign. Thus, stationing the electric vehicle at that location
for daily fixed point sampling will provide data inter-comparing the air monitoring instrument
data onboard the vehicle with that of the Region 5 instruments. This is also the tentative
location for local meteorology measurements to be conducted.
5.2 Process Measurements
All measurements that will be collected by the two monitoring methods are described in Table
5-5. In addition to the below measurements, forward-facing video of the route will also be
collected during each deployment using an on-board webcam.
Table 5-5. Measurements to be conducted during the mobile monitoring campaign
Measurement
Carbon monoxide (CO)
Sulfur dioxide (SO2)
Particle number concentration
(size range 5.6-560 nm, 32
channels)
Particle number concentration
(size range 0.5-20 urn, 52
channels)
Black carbon
Longitude and latitude
3D wind speed and direction
Rate
1 s
1 s
1 s
1 s
1-5 s
1 s
1 s
Instrument
Quantum cascade laser (QCL, Aerodyne
Research, Inc.)
Quantum cascade laser (QCL, Aerodyne
Research, Inc.)
Engine Exhaust Particle Sizer (EEPS, Model
3090, TSI, Inc.)
Aerodynamic Particle Sizer (APS, Model
3321, TSI, Inc.)
Single-channel Aethalometer (Magee
Scientific, AE-42)
Global positioning system (Crescent R100,
Hemisphere GPS)
Ultrasonic anemometer (RM Young, Model )
5.3 General Approach and Test Conditions for Each Experimental Phase
This is an ambient monitoring study that will not have experimental phases. The sampling
details are described in Section 6.
6. Sampling Procedures
6.1 Site-Specific Considerations
The primary objective of this monitoring study is to measure in situ the dispersion of rail yard
emissions to surrounding near-rail yard areas. The measurement results are understood to be
site-specific in nature due to the unique building and vegetation topography, which are known
to impact dispersion. Another site-specific feature is the Cicero rail yard emissions spatial
distribution -the within-yard location and strength of multiple emission points (e.g., diesel-
powered cranes, hostler trucks, and switcher locomotives) will affect the resulting spatial
distribution and chemical composition of emissions. All of these factors will be considered in
the interpretation of results.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 16 of 28
The primary site-specific factors that will affect monitoring procedures are local meteorology,
rail yard source activity, and side-road traffic. The researchers are in communication with
BNSF employees, who have already provided information on general activity trends for the
Cicero rail yard. This information was useful in selecting three different monitoring periods for
sampling. The driving route is designed to provide useful data regardless of wind direction, with
transects located on each side of the yard. In addition to the basic information provided by
BNSF which was utilized in selecting the monitoring time periods, BNSF has also agreed to
provide EPA with daily timestamped recordings of container lifts and trucks entering/exiting the
facility. These data will be used to understand the daily activity trends of the rail yard.
Side-road traffic is an important consideration affecting both the monitoring procedures as well
as the ensuing data analysis. Emissions from local traffic can lead to biases in the monitoring
data and obscure the characterization of the rail yard emissions impact on surrounding areas.
Several methods are used to minimize this potential bias in the data set. First, the site selection
process included surrounding sources, including major roads, as a factor in selecting the Cicero
rail yard site. Second, the sampling timeframes are selected outside of the commute period,
with 0630-0830 and 1430-1830 avoided. Finally, an effort will be made to detect and flag data
that may have a threat of bias due to an individual vehicle's exhaust, which is described in
further detail below.
During sampling, webcams will be used onboard the GMAP to record local vehicle exhaust
episodes. In addition, the operator will also write in the sampling notes (street name and time)
potential exhaust events that were observed and potentially not captured on the webcam video.
Also, the GMAP vehicle will be driven, with safety as the first priority, at a distance behind other
vehicles on the road. Finally, post-processing will be completed of the electric vehicle data set -
an algorithm will be applied that detects and flags time periods with apparent local exhaust
impact, characterized by a sudden sharp spike in carbon monoxide concentrations. This
algorithm will be applied equally to all data recorded by the GMAP vehicle. This algorithm was
developed by Gayle Hagler based upon near-road field studies (Hagler et al., 2010) in the
Triangle Region of North Carolina and found to successfully remove incidences of local
exhaust, with a total loss of data around 2-3% for driving routes with relatively low side road
traffic. The MATLAB code for this algorithm is provided in Appendix A. This algorithm will be
tested again for the CIRYS field study by comparing the flagged time periods for one complete
field sampling deployment with a webcam video record from the dashboard of the GMAP
vehicle. If found to be insufficient in detecting biases, the algorithm will be modified and
rationale for any changes documented. One possibility, specific to this site, is that using carbon
monoxide as the main indicator may not be sufficient to detect truck exhaust. If evaluation with
webcam data reveals that this is the case, black carbon or ultrafine particles would be tested to
see if they are more sensitive indicators of vehicle exhaust.
6.2 Sampling Equipment and Procedures
The sampling equipment used in this study includes one mobile monitoring vehicle equipped
with air monitoring analyzers, GPS, and a webcam, as well as a portable meteorology station.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 17 of 28
Supporting equipment includes a truck equipped with a car-hauling trailer to transport the
electric vehicle as needed.
The following general tasks will be completed as part of the daily instrument deployment
process. Specific details regarding the operation of the GMAP vehicle are found in Appendix B.
1. GMAP electric vehicle will be fully charged (refer to Appendix B, Section 3, GMAP MOP#1)
and equipped with a GPS, webcam, and high time resolution air monitoring instruments.
Time of response will be tested for the various instruments to exactly time-align data.
2. QC checks on all GMAP analyzers will be performed daily during the study.. Prior to
initiating a daily driving period, the multiple computers used for data logging will be time-
aligned with the GPS-derived "true" timestamp.
3. The GMAP electric vehicle will be transported to the field site using a vehicle equipped with
a car-hauling trailer unless a storage and charging area is found near the sampling site.
4. Ambient mobile monitoring sampling will take place for approximately 3 hours. The GMAP
electric vehicle will be driven repeatedly around an assigned route. After approximately 3
hours, the GMAP vehicle will be parked at the location indicated in Figure 5-3 and sampling
will ensue for approximately 1-2 hours, depending on the vehicle battery life.
5. The GMAP air monitoring instrumentation will be shut down and the vehicle will be
relocated to a secure and temperature-controlled environment for recharging and overnight
storage.
The manuals providing procedures for operating the electric vehicle, air monitoring instruments,
and supporting instrumentation are provided in the Appendices B-H as follows:
• Appendix B: GMAP manual
• Appendix C: Hemisphere GPS manual
• Appendix D: EEPS manual
• Appendix E: APS manual
• Appendix F: Aethalometer manual
• Appendix G: QC Laser system technical information
• Appendix H: Ultrasonic anemometer manual
Log sheets will be kept to record the daily sampling events and QA checks - an example log sheet
for the mobile monitoring is below.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 18 of 28
Example:
GMAP Daily Log Sheet
Date: 11/1/09
Sampling Day: 1
Site: Cicero Rail Yard
Operator: James
Sync computer clocks with GPS 0
Fill QCL with liquid N2
PM instrument zero checks:
QCL CO Check:
Aethalometer
EEPS
APS
Time:
7:49 AM
BC:
Total PN:
Total PN:
Time:
11:30 AM
BC:
Total PN:
Total PN:
CO
SO2
Time:
8:05 AM
176ppb
110 ppb
Time:
1:05 PM
164 ppb
115 ppb
QC checks acceptable?
Aethalometer 0
EEPS 0 APS 0 CO 0
S02 0
Driving route
Webcam
Stationary
Start time:
9:15 AM
9:13 AM
11:1 3 AM
End time:
11:13 AM
12:00 PM
12:00 PM
Total number of laps for the run: 5
Observed weather: Light winds from the SW
Comments:
For QCL CO and SO2 Check, using N9 cylinder with 0.18 ppm CO and 0.1 ppm SO9
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 19 of 28
6.3 Quality Control in Sample Analysis
No physical samples will be collected and analyzed - data will be collected using air monitoring
instruments. The Quality Control procedures for the air monitoring and supporting
measurements are described in Section 8.
6.4 Sample Preservation
No physical samples will be collected - data files labeling and storage are discussed in section
6.5.
6.5 Sample Numbering
Air monitoring data will be timestamped and will not have specific sample numbers assigned.
For the GMAP samplers, the individual data files will be uniquely labeled by sampling vehicle,
location, instrument, and date - Location_PlatformName_lnstrument_YYYYMMDD. For
example, UFPs measured using the EEPS on-board the GMAP electric vehicle on November 1,
2010 during the CIRYS Study would be labeled as - CIRYS_GMAP_EEPS_20101101. The
raw data files will maintain this naming scheme in a master database stored by the EPA Project
Leader. Any periods of missing data due to equipment malfunction, severe weather, or
unacceptable quality of data will be documented in the project notebook or electronic files.
6.6 Sample Chain-of-Custody
The original data files will be collected and maintained by ARCADIS personnel. At the
completion of the entire field campaign and data post-processing, final data files and site notes
will be sent to the EPA Technical/Project Leader, Gayle Hagler, for final storage on an EPA
server. An overview of the raw data collection and storage is provided in Figure 6-1. Prior to
sampling with the GMAP vehicle, the on-board data logging computers are manually time-
synchronized to the satellite-based time recorded by the on-board GPS. The instruments log
data using either generic programs (e.g., WinWedge or HyperTerminal) or instrument-specific
programs (e.g., Aerosol Instrument Manager forthe EEPS). Prior experience using these
instruments guides the number of external computers needed to simultaneously log all data
streams or reliance upon internal memory for certain instruments. The GMAP instruments
(GPS, EEPS, APS, AE42, QC laser, and webcam) will log to an external onboard computer and
the meteorology station measurements will also log to a separate and time-synchonized
computer.
After data is recorded and downloaded from the instruments or external data-logging
computers, the data is transferred using a USB drive and a copy is retained by ARCADIS
personnel. Along with field notes recorded electronically, the raw data is later transferred to the
EPA network, with the exception of video files (each about 1 GB) that are stored to an external
hard drive and maintained by the EPA Technical Leader. The raw data files are stored in a
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 20 of 28
folder labeled "raw data" and remain unchanged, with copies of these files made for post-
processing activities. Secondary processing of data, for purposes of aligning real-time
concentrations and location data as well as analysis of trends, is described in detail in Section
9.
Global
positioning
system
Instrument
1
Instrument
2
Instrument
3
Data-logging computers
time-synchronized to CPS
(GPS = true time)
Data continuously
recorded using
Raw data transferred to
external drive and
stored on EPA Server
Large (>1 GB) video or
raw spectral files saved
on external hard drive.
stored by EPA TL
Figure 6-1. Data collection and storage process
7. Measurement Procedures
No analytical methods will be performed on this project, all data is acquired real-time. The critical
measurements for the project were listed in Table 5-5.
7.1 GMAP monitoring
This study will utilize a mobile monitoring vehicle (GMAP) operating in driving-mode
sampling. An image of the GMAP vehicle is provided in Figure 5-1.
Figure 7-1. Mobile monitoring vehicle planned for use in this study.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 21 of 28
While most sampling approaches used in typical stationary fixed site sampling studies
directly translate to this project, such as the use of electrically conductive tubing to minimize
particle loss and careful time-alignment of data-logging laptops, some unique
considerations need to be made for the GMAP vehicle which operates in driving mode.
The two primary additional considerations made are the sampling inlet and the
determination of lag time for sampling instruments.
The GMAP sampling inlet is designed to provide isokinetic conditions while the vehicle is in
motion. Isokinetic conditions are most important for the larger particle sizes (e.g., PM10)
and are generally negligible for gases and ultrafine particles. Two air velocity parameters
are taken into consideration in the design of the inlet -the combined inlet volume flow of
the instruments on-board the electric car and the air flow rate as the vehicle is in motion.
The inlet design assumes an air flow rate for a vehicle driving at approximately 30 mph. In
order to determine whether speed-based correction will be needed for higher driving
speeds, preliminary field tests will be conducted driving the GMAP vehicle over a range of
speeds (0-45 mph) on roads with minimal traffic (refer to Appendix B for more information).
In order to precisely align position and air concentration data for the GMAP vehicle, another
important factor is characterizing the amount of lag time associated with an air sample
transporting through the sample line and measurement by a given air monitoring
instrument. This lag time will be experimentally determined by inducing a sudden
concentration change for the analyte of interest, such as using a HEPA filter for the
particulate instruments, and observing the amount of time before the concentration is
recorded by the monitoring instrument. Further details are provided in Appendix B.
7.2 Calibration Procedures
All equipment will be calibrated annually and/or cal-checked as part of standard operating
procedures. Calibration records are kept on file. Maintenance records are kept for any
equipment adjustments or repairs in project logbooks that include the date and description of
maintenance performed. Details on the instrument-specific calibration and cal-check
procedures are available in the Appendices and Section 8.1.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 22 of 28
8. Quality Metrics
8.1 QC Checks
The QC checks used in the field to assess the QA Objectives (section 8.2) are provided in
Table 8-1.
Table 8-1. Procedures Used to Assess QA Objectives
Measurement
Parameter
Participate size and
number concentration
Participate size and
number concentration
Black carbon
SO2
CO
Ambient wind speed
and wind direction
Location (longitude,
latitude, elevation)
Leaf area index
Analysis Method
EEPS 3090
APS Model 3321
Aethalometer
QC Laser, Aerodyne
QC Laser, Aerodyne
RM Young Ultrasonic
Anemometer Model 81000
Hemisphere GPS
LAI2000
Assessment Method
Single point flow check prior to field campaign
and weekly during campaign;
Daily zero check
Single point flow check prior to field campaign
and weekly during campaign;
Daily zero check
Single point flow check prior to field campaign
and weekly during campaign;
Daily zero check
Pre-deployment multi-point calibration check
Daily zero/span check
Pre-deployment multi-point calibration check
Daily zero/span check
Pre-deployment calibration by Metrology Lab
Pre-deployment comparison of measured GPS
data with known reference location
Comparison of LAI data with known range of
historically observed values for similar vegetation
types.
8.1.1 Particle Measurement Instrument Assessment
The EEPS, APS, and Aethalometer measure particulate components based on
manufacturer calibration. For the instruments used in this study, the manufacturer
calibration took place in 2009 for the EEPS and APS. The Aethalometer was calibrated in
2008. To test each instrument's performance prior to use in the field and periodically during
the sampling campaign, single-point flow verification and a zero check should be
conducted. The flow verification should be conducted using a calibrated flow meter and
flows should be within 10% of the set point. The zero check is conducted by attaching a
high efficiency particulate air (HEPA) filter to the sampling inlet, which removes >99% of
particulates of diameter >0.3 urn. While the HEPA filter is in place, downstream particulate
concentrations should read near zero for the instrument to be deemed acceptable. The
zero check should be performed prior to each daily deployment to the field and a flow
check should be performed weekly during the field campaign. Given a failure in meeting
this data quality indicator, response actions include, but are not limited to, (1) performing
cleaning maintenance, (2) changing sampling inlet, and (3) seeking technical support from
the instrument manufacturer.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 23 of 28
8.1.2 Quantum Cascade Laser assessment
The CO and SO2 data from the QCL will be verified against gas standards which bracket
the anticipated range of ambient concentrations. Given a failure to meet this data quality
indicator, response actions will include, (1) verifying the data collection process is being
performed correctly, (2) seeking technical support from the instrument manufacturer.
8.1.3 Global Positioning System Assessment
The GPS system will be verified by driving along a specific route and remaining stationary
at a known location, then comparing reported longitude/latitude against mapping data.
Several software or internet-based programs are available to determine whether reported
data matches the actual route, including ArcGIS, MATLAB, and Google Earth Pro.
8.1.4 Ultrasonic Anemometer Assessment
The ultrasonic anemometer DQIs are checked annually as part of a routine calibration
procedure. The specific ultrasonic anemometer used for this study has recently undergone
repair and recalibration by the manufacturer (Summer 2010).
8.2 QA Objectives and Acceptance Criteria
The Data Quality Indicator goals for accuracy, precision, and completeness for this project are
listed in Table 8-2. Any failure of the instrumentation to meet the DQI goals will be reported to
the EPA TIP. Data collected during time periods in non-attainment with DQI goals will be
flagged as questionable, but not necessarily considered invalid. Corrective action to be taken
depends on the nature of the problem encountered. For example, given an error in sampling
flow, careful cleaning of the inlet may be required.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 24 of 28
Table 8-2. Data Quality Indicator Goals for the Project
Measurement
Parameter
Carbon
Monoxide
Sulfur dioxide
Black carbon
Particulate size
and number
concentration
Particulate size
and number
concentration
Wind speed and
direction
Location
Analysis
Method
Quantum
Cascade
Laser
Quantum
Cascade
Laser
Aethalo-
meter
EEPS 3090
APS Model
3321
Ultrasonic
anemometer
Hemisphere
GPS
Assessment
(1) Initial flow check
(2) Zero/Span check
in the field
(1) Initial flow check
(2) Zero/Span check
in the field
(1) Initial single point
flow check
(2) Daily zero check
(1) Initial single point
flow check
(2) Daily zero check
(1) Initial single point
flow check
(2) Daily zero check
Data will be
compared with field
operator observations
on log sheet.
(1) Status lights
indicate
collected signal
(2) Review of
mapped location
with pre-
designated route
and stationary
location.
Criteria
Accuracy
Within +/-
20% of
calibration gas
Within +/-
20% of
calibration gas
(1)+/-10%of
set-point
(2) 5-min
average at
<20% of
ambient3
(1)+/-10%of
set-point
(2) 5-min
average at
<20% of
ambient3
(1)+/-10%of
set-point
(2) 5-min
average at
<20% of
ambient3
General
matching of
wind direction
and speed
+/- 1 0 m of a
known
location
Complete
-ness
90%
90%
90%
90%
90%
QD%
95%
Precision13
+/-10%
+/-10%
N/A
N/A
N/A
N/A
+/-10m of
a known
location
Corrective Actions Given Failure to meet Criteria
(1) Inlet will be checked for flow obstructions
(2) Instrument troubleshooting will take place.
(1) Inlet will be checked for flow obstructions
(2) Instrument troubleshooting will take place.
(1) Sampling inlet will be checked for obstructions. If flow
errors continue, instrument troubleshooting and/or flow
recalibration will take place.
(2) Instrument connections will be checked and zero-check
repeated. Data collection will continue given repeat failure,
but data will be flagged.
(1) Sampling inlet will be checked for obstructions. If flow
errors continue, instrument troubleshooting and/or flow
recalibration will take place.
(2) Instrument connections will be checked and zero-check
repeated. Data collection will continue given repeat failure,
but data will be flagged.
(1) Sampling inlet will be checked for obstructions. If flow
errors continue, instrument troubleshooting and/or flow
recalibration will take place.
(2) Instrument connections will be checked and zero-check
repeated. Data collection will continue given repeat failure,
but data will be flagged.
(1 ) Orientation of sonic anemometer will be checked and
corrected if found to be out of alignment.
Sampling will discontinue until GPS is determined to be
functioning properly.
3The HEPA filter removes >99% of particulates of diameter >0.3 jjm, however the particle instruments also measure particles under 0.3 j^m, which may have a higher penetration efficiency.
""Precision will be based upon comparison with a gas standard. GPS data precision will be compared by looking at the variability of the location data when the car remains parked at a
single location.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 25 of 28
9. Data Analysis, Interpretation, and Management
9.1 Data Reporting
Research results are intended for publication in scientific journals, thus no writing of internal
EPA reports is expected. The EPA Technical Lead, Gayle Hagler, will be responsible for
generating a data report for internal use among EPA scientists. This report will include
information on the sampling collection times, field notes, and preliminary data review
(completeness, QC checks).
9.2 Data Validation
Verification and validation of the procedures used to collect and analyze data are critical to
achieving project objectives. Data validation for this study will be accomplished through a
review of quality control checks conducted daily for the instrumentation as described in Table 8-
2. This review will determine whether or not instrumentation had acceptable performance and
the data useable in analysis. ARCADIS will be responsible for the operation all field
instrumentation. The EPA Technical Lead, Gayle Hagler, will be responsible for reviewing the
GMAP vehicle data.
9.3 Data Analysis
Following the collection of raw data, as described in Section 6, the GMAP data are processed
using several standard algorithms developed in MATLAB, which is described in Figure 9-1 - (1)
Adjustment for lag time, (2) Combining GPS location data and air monitoring data into a joint
matrix that is now time and spatially-resolved concentration data, and (3) Spatially consolidating
data from repeat drives into spatial increments of interest (user-defined) for purpose of
calculating averages or other statistics. Data analysis can take place at various levels of post-
processing. For example, inter-comparing data for the same variable (e.g., black carbon) may
only require that the data be time-aligned (step 1). Observing concentration changes in both
time and location would require steps 1 and 2 to be conducted. Finally, looking at 2-hour
average concentration maps would require steps 1-3 to be completed. While the algorithms
used to process steps 1-3 are customized based upon each specific instrument's data, the
common algorithm elements are provided in Appendix H.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 26 of 28
Raw data files
Import Into data-
processing program
(e.g.. MATLAB) and
run algorithms
{2} Combine GPS
time series and air
monitoring data
time series
{3} Consolidate data from
repeat drives at defined
spatial Increment (e.g. 10
m averages)
longilude
ongrtiide
Secondary variables used in analysis
Figure 9-1. Post-processing of mobile monitoring raw data for use in analysis.
For this study, data analysis after the above processing steps will include parallel time series
and correlation analysis of the air pollutant measurements (following step 1) as well as
geospatial and temporal analysis (following step 2 and/or step 3) of the driving-mode mobile
monitoring vehicle data. These and other analyses may lead to further post-processing of data,
dependent on project needs. Additional data used for interpretation will include regional
meteorology data and other air pollutants measured simultaneously on-board the electric
vehicle. In addition, Region 5 is planning to set-up a fixed monitoring site adjacent to the rail
yard and may have a number of continuous measurements that duplicate those in the electric
car. This monitoring site will have a separate QA document processed by Region 5 QA staff. If
this fixed monitoring data is available concurrent with the mobile monitoring study, this will be a
second data set available for analysis and will provide context for the mobile monitoring data.
Gayle Hagler will be the main individual responsible for the analysis of the GMAP data and any
combined analysis of the GMAP and Region 5 fixed monitoring site data.
In order to meet project objectives of characterizing whether or not, and to what extent, rail
yard-related air pollutants are elevated over background concentrations, it is important to define
how "background" is defined for this case. The concept of "background" ranges in the scientific
community to the natural ambient background, without any development; to a rural setting,
removed from any major industry; to an "urban background", in a developed area but removed
from any major source. For this study, the urban background concept will be applied and
defined as upwind-of-rail-yard areas that are greater than 200 m from the road, with minimal
traffic on roadways monitored. The lower wind threshold will be 1 m/s - if wind speeds are
lower, the meteorology conditions will be considered low speed, mixed winds and an upwind
area will not be defined. For days with winds > 1 m/s, the mobile monitoring data covering
urban background areas meeting the aforementioned criteria will be averaged and the
variability of the background will be quantified. These upwind data will be compared vis-a-vis
data covering areas downwind of the rail yard, again on roadways with minimal traffic, to
evaluate whether downwind concentrations are higher than urban background conditions.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 27 of 28
9.4 Data Storage Requirements
No physical samples will be collected or require storage. Section 6.6 discusses the chain-of-
custody and storage for the mobile monitoring data.
10. Reporting
10.1 Deliverables
Deliverables from this study include final quality-assured field data and manuscripts for
publication in scientific journals.
10.2 Expected Final Products
Anticipated final products for this study are peer-reviewed, published research papers in
science journals and presentations at scientific conferences.
11. References
Baldauf, R., Thoma, E., Isakov, V., Long, T., Weinstein, J., Gilmour, I., Cho, S., Khlystov, A.,
Chen, F., Kinsey, J., Hays, M., Seila, R., Snow, R., Shores, R., Olson, D., Gullett, B., Kimbrough,
S., Watkins, N., Rowley, P., Bang, J., and Costa, D. 2008a. Traffic and meteorological impacts on
near road air quality: Summary of methods and trends from the Raleigh Near Road Study.
Journal of the Air & Waste Management Association 58:865-878.
Baldauf, R., Thoma, E., Khlystov, A., Isakov, V., Bowker, G., Long, T., Snow, R., 2008b. Impacts
of noise barriers on near-road air quality. Atmospheric Environment 42, 7502-7507.
Chang, 2007. Roseville Railyard Air Monitoring Project Data Analysis Report. Placer County Air
Pollution Control District Board of Directors Meeting, Agenda Date: April 12, 2007. Website link:
http://www. placer, ca. gov/Departments/Air/~/media/apc/documents/UP/UPRRProjectReport2ndYear
%20pdf.ashx.
Hagler, G.S.W., Baldauf, R.W., Thoma, E.D., Long, T.R., Snow, R.F., Kinsey, J.S., Oudejans, L.,
and Gullett, B.K., 2009. Ultrafine particles near a major roadway in Raleigh, North Carolina:
Downwind attenuation and correlation with traffic related pollutants. Atmospheric Environment
43:1229-1234.
Hagler, G.S.W., Thoma, E.D., and Baldauf, R.W., 2010. High-resolution mobile monitoring of
carbon monoxide and ultrafine particle concentrations in a near-road environment. Journal of the
Air & Waste Management Association 60:328-336.
Heiken, J.G., 2009. Rougemere Rail Yard Emission Inventory; prepared by Sierra Research, Inc.,
Sacramento, CA, for the Lake Michigan Air Directors Consortium. Report is available upon
request by the TLP.
Heist, O.K., Perry, S.G., and Brixey, L.A., 2009. A wind tunnel study of the effect of roadway
configurations on the dispersion of traffic-related pollution, Atmospheric Environment 43:5101-
5111.
-------
APPCD QAPP - Cicero Rail Yard Study
Revision 1
January 13, 2011
Page 28 of 28
Hu, S., Fruin, S., Kozawa, K., Mara, S., Paulson, S.E., Winer, A.M., 2009. A wide area of air
pollutant impact downwind of a freeway during pre-sunrise hours, Atmospheric Environment,
43:2541-2549, doi:10.1016/j.atmosenv.2009.02.033
Thoma, E.D., Shores, R.C., Isakov, V., Baldauf, R.W., 2008. Characterization of near road
pollutant gradients using path-integrated optical remote sensing. Journal of the Air and Waste
Management Association, 58:879-890.
Turner, J.R., Yadav, V., and Feinberg, S.N., 2009. Data Analysis and Dispersion Modeling of the
Midwest Rail Study (Phase I) - Final Report. Available at:
http://www.ladco.org/re ports/genera l/new_docs/WUSTL_MidwestRailStudy_FinalReport.pdf
Westerdahl, D., Fruin, S.A., Fine, P.M. and Sioutas, C., 2008. The Los Angeles International Airport
as a source of ultrafine particles and other pollutants to nearby communities, Atmospheric
Environment, 42:3143-3155, doi:10.1016/j.atmosenv.2007.09.006
Zhu, Y. F., W. C. Hinds, S. Kim, and Sioutas, C., 2002. Concentration and size distribution of
ultrafine particles near a major highway. Journal of the Air & Waste Management Association
52:1032-1042.
-------
Quality Assurance Project Plan (QAPP)
US EPA Region 5
Regional Applied Research Effort (RARE) Project
Cicero Rail Yard Study (CIRYS)
Stationary Special Purpose Monitoring (SPM) of Near Rail
Yard Air Quality
Version 1.0
Approval Date: *® 212010
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 2 of 15
A, PROJECT MANAGEMENT
Al. Approval Sheet
Cicero Rail Yard Study (CIRYS) - Stationary Special Purpose Monitoring (SPM) of Near Rail Yard Air
Quality
Monica Paguia
Region 5 Air & Radiation Division
U.S. Environmental Protection Agency
Project Leader/QA Coordinator
Date
Chad McEv'oy* j
Region 5 Air & Radiation Division
U.S. Environmental Protection Agency
Project Scientist
mo
Date
Loretta Lehrman Date
Region 5 Air & Radiation Division
U.S. Environmental Protection Agency
Quality Assurance Manager/Section Chief
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 3 of 15
A2. Table of Contents
A. PROJECT MANAGEMENT
Al. Approval Sheet p. 2
A2. Table of Contents : p. 3
A3. Distribution List p. 4
A4. Project Organization p. 5
A5. Problem Definition & Background p. 6
A6. Project/Task Description p. 6
A.7 Quality Objectives and Criteria p. 8
AS. Special Training/Certification p. 9
A9. Documentation and Records p. 9
8. DATA GENERATION & ACQUISITION p. 9
Bl. Sampling Process Design p. 9
B2. Sampling Methods p. 9
B3. Sample Handling and Custody p. 10
B4. Quality Assurance/Quality Control p. 10
B5. Instrument/Equipment Testing, Inspection, and Maintenance p. 11
B6. Instrument Calibration and Frequency p. 11
87. Inspection/Acceptance for Supplies and Consumables p. 12
88. Non-Direct Measurements p. 12
89. Data Management p. 13
C. ASSESSMENT & OVERSIGHT p. 14
Cl. Assessment and Response Actions p. 14
C2. Reports to Management p. 14
D. DATA VALIDATION & USABILITY p. 14
Dl. Data Review, Verification, and Validation p. 14
D2. Verification and Validation Methods p. 14
APPENDICES
1. Instrument Manuals
2. SOPs
3. CIRYS - High Resolution Mobile Monitoring of Near Rail Yard Air Pollution QAPP
4. 40 CFR Appendix A
5. QA Handbook Vol. IV
6. Example Work Sheets
7. List of Qualifiers
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 4 of IS
A3. Distribution List
The QAPP wilt be distributed electronically. A hardcopy will be kept at the fixed site location.
EPA Region 5
Loretta Lehrman
Monica Paguia
Chad McEvoy
Michael Compher
Jaime Wagner
Basim Dihu
Anthony Ross
Jesse McGrath
EPA ORD
Gayte Hagler
Eben Thoma
BNSF
Pay) Nowicki
David Seep
Michael Stanfitt
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page S of 15
A4. Project Organization
Table A-l: Key points of contact
Name
Monica
Paguia
Chad
McEvoy
Loretta
Lehrman
Gayle
Hagler
Eben
Thoma
Organization
Affiliation
EPA Region 5
EPA Region 5
EPA Region 5
EPA/NRMRL
EPA/NRMRl
Title
Project
Leader/QA
Coordinator
Project
Scientist
Supervisor/
QA Manager
Project team
member
Project team
member
Responsibilities
Overall project leadership for
study; data review, verification,
and validation; and maintaining
official QAPP
Project coordination. Lead for
fixed site technical activities and
data generation.
Project oversight and Quality
Assurance
ORD technical leadership for
mobile monitoring, data analysis.
ORD technical support for mobile
monitoring
Contact Information
Phone: (312) 353-1166
Email:
paguia, monica@epa.gov
Phone: (312) 886-6084
Email:
mcevoy.chad@epa.gov
Phone: (312) 886-5482
Email:
lehrman.loretta@epa.gov
Phone: (919) 541-2827
Email:
hagler.gayle@epa.gov
Phone: (919) 541-7969
Email:
thoma.eben@epa.gov
Table A-2: Additional Project Team Members
Name
Michael
Compher
Jaime
Wagner
Basim Dihu
Jesse
McGrath
Anthony
Ross
Organization
Affiliation
EPA Region 5
EPA Region 5
EPA Region 5
EPA Region 5
EPA Region 5
Title
Project team
member
Project team
member
Project team
member
Project team
member
Project team
member
Responsibilities
Project coordination
Field Study
Communications
Field study support
Field study support
Field study support
Contact Information
Phone: (312) 886-5745
Email:
compher.michael@epa.gov
Phone: (312) 886-9402
Email: wagner.jaime@epa.gov
Phone: (312) 886-6242
Email: dihu.basim@epa.gov
Phone: (312) 886-1532
Email: mcgrath.jesse@epa.gov
Phone: (312) 353-0826
Email: ross.anthony@epa.gov
-------
C1RYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1,0
10/21/10
Page 6 of 15
AS. Problem Definition & Background
This research study was originated by an EPA Region 5 RARE proposal, which hypothesized that rail
yard emissions may locally elevate air pollutant levels and contribute to regional PM2,5
nonattainment status. The initial intent of this study was to better understand carbon emissions
near rail yards in Region 5, determine how locomotive emissions contribute to local PM2.5
concentrations, and to further the overall objective of better understanding rail yard emission
impacts on PM2.5 concentrations and other pollutants in the ambient air within Region 5.
Phase I of the RARE study focused on the CSX Rougemere Rail yard in Dearborn, Michigan in a multi-
component research effort that included a rail yard emissions inventory, dispersion modeling, and a
several month monitoring campaign. This campaign measured carbonaceous particulate matter
(black carbon, elemental carbon, and organic carbon) at two sites immediately adjacent to the rail
yard and additional site located within a nearby residential area which represented urban
background concentrations. A formal technical report on Phase I results has been completed. The
results documented a reduction in annual PM2.S emissions due to locomotive replacements and fuel
changes which occurred at the rail yard and it recommended that characterization of other nearby
sources be performed. Because our research objective is to better our understanding of rail yard
emission impacts within Region 5, the decision was made to continue Phase II at a larger rail yard
within an urban, as opposed to an industrial setting.
This QAPP applies to the Phase II of this study; particularly the special purpose stationary monitoring
site. The primary objective of Phase II, which is planned to occur near the BNSF Cicero Rail Yard in
Cicero, It is to characterize the spatial extent and variability of near-rail yard impact on air pollutant
concentrations in an urban area.
Region 5 will support the stationary site. It is located near- downwind of the Cicero Rail Yard and
along the mobile monitoring route. In addition to providing context for the mobile monitoring data,
it will provide a continuous time series of meteorology and concentrations of fine particulate matter
(PM2.5), carbon monoxide (CO), sulfur dioxide (SO2), black carbon (BC), and nitrogen oxides (NOx).
A separate component of Phase II uses a mobile monitoring approach to measure concentrations
of multiple species (CO, SO2, BC, and size-resolved particle number concentration) in real-time while
driving on multiple transects on upwind and downwind sides of the rail yard. This other component
is addressed in a separate QAPP authored by EPA's Office of Research and Development (ORD).
A6. Project/Task Description
This QAPP addresses the special purpose stationary monitoring component of Phase II and involves
three major tasks.
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 7 of IS
1. Special Purpose Monitoring: Install, operate, and maintain continuous measurements for PM2 s<
BC, S(>2,CO, NOx and meteorological data. This monitoring is planned to start from mid-October
2010 and continue for six months to a year.
The site location is downwind (based on prevailing winds in the Cicero area), just northeast of
the Cicero Rail Yard as shown in figure A-l as indicated by the orange circle. It was chosen
based on prevailing wind direction, openness (no obstructions to the site), ease of access,
security, and electrical availability.
2. Data Analysis: Prepare a database with the measurements from the special purpose monitoring
study, perform data validation, and conduct data analyses to characterize the spatial extent and
variability of near-rail yard impact on air pollutant concentrations and to provide context for the
data generated from the mobile monitoring vehicle. Data validation and preliminary analyses
will occur as data is being collected and throughout the project period. Data analysis will
include assessing concentrations as a function of wind direction, wind speed, time of day, and
day of the week. In addition, the fixed site data trends will be compared with that of the mobile
monitoring vehicle during the mobile monitoring time periods. Data analysis will continue for
several months after sampling.
3. Reporting: Develop presentations and prepare an EPA internal technical memorandum
summarizing the monitoring field study efforts and data analysis performed by Region 5. The
data collected from this fixed site may be used by ORD for research purposes and published in
scientific journals.
Figure A-l Satellite view of Cicero Rail Yard and the fixed site location
-------
C1RYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 8 of 15
A.7 Quality Objectives and Criteria
The primary objective of this research study is to characterize the spatial extent and variability of
near-rail yard impact on air pollutant concentrations in an urban area. Because the special purpose
monitoring data collected at the fixed site will be used together with the mobile monitoring data
and possibly other nearby regulatory ambient air monitoring data, data quality objectives must
meet similar or more rigorous requirements.
The location of the fixed monitoring site allows for samples to be representative of near- rail yard
ambient air concentrations. The site Is located to the northeast of the rail yard, which is in the
prevailing wind direction and thus is frequently expected to be the receptor of transported rail yard
emissions. For a given wind direction and wind speed, the measured concentrations will be a
function of the emissions strength and emission location for a particular species throughout the rail
yard. Thus, the concentrations measured adjacent to the rail yard may be more heavily impacted by
emissions in the near-field zone. This would be true of any near-rail yard monitoring location, thus
this site is considered to be generally representative of near-rail yard air quality.
• Precision is a measurement of mutual agreement between two measurements of the same
property usually under prescribed similar conditions. At a minimum, bi-weekly QC checks
will be conducted on the gaseous analyzers. The met equipment has recently been
recertified.
• Bias (a combination of precision and bias) Is defined as the systematic or persistent
distortion of a measurement process which causes errors in one direction. Audits will be
performed on the gaseous analyzers and flow checks conducted on the PM2.S sampler &
aethalometer. The mobile (GMAP) monitoring data will be collected simultaneously with
the special purpose monitoring data at the same location of the fixed site for each day the
GMAP is deployed.
• Completeness refers to a measure of the amount of valid data obtained from a
measurement system compared to the expected amount obtained under normal conditions.
A data completeness goal of at least 75% is expected.
• Sensitivity or Detect ability refers to the low critical range value that a method could
reliably discern.
Table A-3 Method Detection limits {MDL)
Measurement
Parameter
PM2.S
Black Carbon
Analysis Method
EBAM (beta attenuation method)
Optical Transmission Method
MDL
Dependant on time
resolution
Dependant on flow
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1,0
10/21/10
Page 9 of 15
NOx
S02
CO
Ambient wind speed
TECO 42
TEI 450C, API 101A, & API 101E
API 300E
Met One
rate and time
resolution
0.5 ppb
0.5 ppb
40 ppb
0.3 m/s
A8. Special Training/Certification
All personnel have the appropriate training and experience necessary to fulfill their role and
responsibilities needed to implement this project and meet the required quality objectives.
A9. Documentation and Records
Bound logbooks will be maintained at the stationary monitoring site. Information documented in
these logbooks will include at least the date, time, site operator initials, calibrations and audits
conducted, preventative maintenance, and troubleshooting performed on any of the instruments,
along with any other pertinent observations and information.
Additionally, a binder containing all QC documentation (audit sheets, calibration sheets and
standards certification certificates) will be maintained by U.S. EPA.
B. DATA GENERATION & ACQUISITION
This project will rely on this QAPP, EPA Standard Operating Procedures (SOPs), and manufacturer's
instrument manuals to acquire and generate data, instrument manuals and SOPS can be found in
the appendices.
Bl. Sampling Process Design
The sampling process is designed to help better understand the local impact of rail yard emissions
on ambient air pollutant concentrations, to characterize the air quality in this area over a longer
period of time than the mobile monitoring campaign, and to provide supporting information for the
mobile monitoring effort. The sampling approach consists of continuous monitoring of PM2.5 BC,
SO2, CO, NOx, wind speed/wind direction and other meteorological data at a fixed location over a six
month to one year time period.
B2. Sampling Methods
All data will be collected using automatic continuous air pollution measurement analyzers and
meteorological sensors. Table A-3 lists the instrument methods used. Neither manual sampling nor
analytical methods (sample analyses) are needed for this project.
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 10 of 15
B3. Sample Handling and Custody
Sample handling is not necessary for this project. The used tape from the continuous PM2.smonitor
and aethalometer will be properly labeled, transferred, and stored at the Region 5 Air Monitoring
Laboratory. This will be appropriately documented; no formal chain of custody is necessary.
84. Quality Assurance/Quality Control
The purpose of Quality Control (QC) is to establish confidence, demonstrate reliability, and ensure a
sufficient level of data quality for its intended use. The appropriate quality control checks will be
performed on all instruments prior to the start and at the end of the project and according to the
schedule as shown in Table 6-1, Independent audits and flow checks will be performed by an
outside source.
See Tables B-l, B-2 and section B6 for detailed information regarding QC for this project.
Table B-l. Procedures Used to Assess QA Objectives
Measurement
Parameter
PM2.5
Black Carbon
NOx
S02
CO
Ambient wind speed
and wind direction
Assessment Method
Flow Checks; flow check
using independent standard
Flow Checks
Single point QC Check (zero
& span);
Audit (3 consecutive levels at
80% of measured cone.)
Single point QC Check (zero
& span);
Audit (3 consecutive levels at
80% of measured cone.)
Single point QC Check (zero
& span);
Audit (3 consecutive levels at
80% of measured cone.)
Certification;
General Inspection and
maintenance
Minimum
Frequency
Monthly; every 6
months
Monthly
Every 2 weeks;
Every 6 months
Every 2 weeks;
Every 6 months
Every 2 weeks;
Every 6 months
Pre-deployment;
Every 6 months
Acceptance Criteria
<4% of Standard; ±5%
of Design Value
5*10%
£10%; <1S%
£10%; <15%
S10%; <15%
See manual
-------
C1RYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 11 of 15
65, Instrument/Equipment Testing, Inspection, and Maintenance
All instruments will be inspected and tested prior to deployment and installment at the fixed site.
General inspection of the instruments will be conducted during each visit to ensure proper
performance. Preventative maintenance and troubleshooting will occur on an as needed basis and
documented accordingly.
Quality control for the meteorological sensors is achieved by using certified MET sensors. The
installation of sensors will follow the Quality Assurance Handbook Volume 4 and the instrument
manual.
66. Instrument Calibration and Frequency
Calibrations for the continuous gaseous analyzers use the following methods:
1. Multi-point calibration of the analyzers at the start-up of the project.
2. Audit or verification of the analyzers at the end of the project.
3. Multi-point calibration of the analyzers after any major repairs.
4. Multi-point calibration of the analyzers if a zero/span/precision (audit) check of the analyzer
exceeds +/* 10% of the expected known value.
At a minimum, a calibration sequence will include at least a zero point, a span point of
approximately 80% of the instrument range and at least 3 additional points equally spaced at
intervals of 20% of the range and the span point.
For example, for monitors ranging from 0 - 0.500 ppm the following concentrations could be
generated and introduced into the reporting instrument:
Zero air
0.030 ppm - 0.100 ppm
0.150 ppm - 0.200 ppm
0.250 ppm -0.300 ppm
0.350 ppm - 0.450 ppm
If the lowest expected calibration point cannot be reached (because of calibration system limitations
at low flows) then the lowest possible point will be used.
ALL CALIBRATION POINTS DURING THE CALIBRATION MUST BE WITHIN 3.0% OF THE EXPECTED
VALUE.
Following a successful calibration the converter efficiency will be calculated according to the
manufacturer's manual and should be > 90% and <105%. If not, the converter will be replaced as
soon as possible. If a new converter is unavailable and the instrument has demonstrated acceptable
-------
C1RYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 12 of 15
audit checks then the monitoring location will continue and all data from the instrument will be
flagged accordingly.
A slope and intercept will also be calculated using the standard concentration and the analyzer
response. The slope should be between 1.005 and 0.950 and the intercept should be between -5.0
and 5.0 ppb.
At no time will monitoring data be reported as valid if any point in an audit is over 15% difference
when comparing the standard concentration versus the analyzer response.
A one point calibration will be performed on the continuous PM2 5 analyzer prior to the start of
sampling.
Re-calibration of the aethalometer's flowmeter is only necessary if there is serious reason to believe
that the flowmeter response is incorrect.
All calibration responses, all audit check responses, and any adjustments made will be documented
in the worksheets, checklists, or logbooks.
87. Inspection/Acceptance for Supplies and Consumables
All supplies and consumables are purchased from established vendors to maintain consistent
quality. All calibration gases are NIST traceable and have been purchased from recognized,
reputable suppliers.
An adequate amount of supplies, consumables, and gasses will be purchased prior to the start of
sampling. They will be inspected at least every two months and re-ordered if necessary.
88. Non-Direct Measurements
The calibration gases used for multi-point calibrations and audit checks will be of a certified mixture
type as specified by the "EPA Traceabitity Protocol for Assay and Certification of Gaseous Calibration
Standards (Revised September 1993)". Gases for audits and gases for calibrations will be
independent. These gases will be labeled clearly to indicate whether they are to be used for audits
or for calibration.
A multi-gas dilution calibrator will be used for generating the calibration challenge points for audits
and calibrations. Calibrators for audits and for calibrations will be independent. Both gas blenders
will be labeled clearly to indicate whether they to be used for audits or calibrations. The flow rates
of the multi-gas dilution calibrators will be checked against a NIST traceable flow device at a
frequency of every 6 months (due next in Jan. 2011). If systematic problems are suspected during
any QC activity it is recommended that the calibration system is recertified prior to the diagnosis of
required repairs.
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 13 of 15
Quality control for the dataloggers used to capture and store the data will be achieved by verifying
that the analyzer readings match the dataloggers instantaneous reading responses to within 2.0
ppb. Caution will also be necessary in ensuring and documenting that the analog output channels of
the pollutant analyzers match the reading on the data logger. Adjustments will be made according
to the manufacturer instrument manual to the analyzer analog output channels zero and span
settings if this comparison does not show agreement. All adjustments made will be recorded in the
site logbooks.
Cellular phone times may be used for comparing times on analyzers and data loggers as long as a
brief comparison has been established between the cellular phone and the Official U.S. Time Clock
that can be found online at http://www.time.gOv/timezone.cgi7central/d/-6/iava. A comparison of
analyzer time versus data logger time versus standard time should be completed and documented
at each site visit. Times should be adjusted if found to be off by more than 60 seconds. Any
adjustments and the magnitude of any adjustments will be recorded in the site logbook. This site is
in the central time zone. The time will be recorded in Central Standard Time (CST).
Table B-2. QA/QC Non-Direct Measurements
Parameter
Data Capture
Calibration Gases
Multi-Gas dilution
calibrators
Dataloggers
Times on analyzers
Assessment Method
download
Certification
Checked against NIST traceable flow
device
reading responses within 2.0 ppb
Official U.S. Time Clock
http://www.time.ROv/ttrnezone.CRiPc
e ntxal/d/i6/j3ya .
Minimum Frequency
monthly
NOX - every 24 months
CO - every 36 months
S02 - every 24 months
Every 6 months (due in Jan. 2011)
Every site visit
Every site visit
B9. Data Management
Data will be downloaded from the analyzers and samplers on an as needed basis or at least a
monthly basis; whichever is more frequent. Raw data is archived on EPA Region 5 ARD share drive,
G.
Data will be provided to ORD electronically for comparison to mobile monitoring data, which will be
stored on EPA servers with the mobile monitoring data set.
All data, documentation and records associated with the special purpose monitoring at the fixed
site, will be maintained by the USEPA Region 5 Air Monitoring and Analysis Section (AMAS) for at
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page 14 of 15
least 5 years from the end of the project. These will be stored electronically on the Air & Radiation
(ARD) G drive and hardcopy in AMAS filing cabinet.
C. ASSESSMENT & OVERSIGHT
Cl. Assessment and Response Actions
Precision and accuracy (zero and span checks) measurements are conducted on the gaseous
analyzers every 2 weeks. The flow checks on the continuous PM2.5 sampler and aethalometer are
conducted monthly. A site evaluation will be conducted prior to the start of data collection. A
technical system audit may be conducted within a few weeks of the start of data collection to
ensure all procedures are in place and to ensure the level of QA/QC is sufficient. All audit/check
worksheets, checklists, and reports will be reviewed by the project QA Coordinator on a quarterly
basis.
Any corrective actions needed for any part of the project will be properly verified, implemented and
documented by the QA Coordinator.
C2. Reports to Management
An EPA internal technical memorandum summarizing the special purpose monitoring field study
efforts and data analysis performed by Region 5 will be developed for management after the
sampling and data analysis has been concluded. The data collected from this fixed site may be used
by ORD for scientific research and published in scientific journals and in presentations.
D. DATA VALIDATION & USABILITY
01. Data Review, Verification, and Validation
Data Review - General review of the data will be based on level of data capture or completeness
and by using graphs of the data to determine any trends or anomalies.
Data Verification - Verification includes review of the QC checks meeting the criteria and comparing
them and their implications to the data.
Data Validation - Validation of the data is determined by usability of the data. Any data outliers will
be flagged and appropriately qualified.
D2. Verification and Validation Methods
Every site visit or at least each month, whichever is more frequent, the site operator will review the
sampling logs and raw data. Any deviations from the QAPP will be documented and reported to the
-------
CIRYS - Stationary SPM of Near Rail Yard Air Quality QAPP
Version 1.0
10/21/10
Page IS of IS
project QA Coordinator. Project QA status reports/checklists will be developed quarterly and
submitted to the QA Coordinator for review.
The data generated during this project will be reviewed, verified, and validated by the project QA
coordinator on a quarterly basis by comparison with analyzer and sensor performance parameters
and quality control results.
-------
Cicero Rail Yard Study (CIRYS) - Final Report
Appendix B: Mobile Monitoring Quality Check Results
QC Check Results (red text = flagged) - Check 1
Day
Aethal
EEPS
APS
CO
S02
ometer
DQI (Table 8-2
ofQAPP)
10/27/2010
10/28/2010
10/29/2010
10/30/2010
10/31/2010
11/1/2010
11/3/2010
11/4/2010
11/5/2010
11/6/2010
11/7/2010
11/8/2010
11/10/2010
11/11/2010
11/12/2010
11/13/2010
11/15/2010
11/16/2010
11/17/2010
11/18/2010
11/19/2010
11/20/2010
11/21/2010
aEEPS zero checks
<20%
n/a
14%
5%
43%
110%
0%
2%
9%
7%
4%
2%
2%
1%
12%
8%
9%
2%
11%
20%
10%
3%
16%
6%
are evaluated
<20%
1%
0%
1%
1%
1%
1%
1%
0%
0%
1%
1%
2%
1%
1%
3%
2%
3%
0%
0%
1%
5%
1%
2%
for particles
<20%
3%
0%
0%
0%
0.4%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
>100
80-120%
97%
99%
100%
101%
99%
100%
99%
98%
98%
98%
99%
98%
97%
101%
100%
100%
102%
101%
97%
94%
97%
98%
96%
nm
80-120%
68%
67%
69%
69%
64%
68%
69%
67%
68%
66%
68%
68%
65%
71%
73%
71%
66%
71%
68%
69%
69%
69%
70%
Aethal
ometer
<20%
n/a
9%
-7%
9%
56%
2%
0%
5%
13%
2%
7%
2%
1%
5%
8%
3%
3%
20%
16%
8%
-2%
10%
13%
QC Check Results (red text = flagged) - Check 2
EEPS
<20%
n/a
8%
1%
5%
1%
0%
2%
3%
2%
1%
1%
1%
1%
6%
1%
1%
2%
0%
0%
2%
0%
1%
0%
APS
<20%
n/a
4%
0%
5%
0%
0%
0%
0%
1%
0%
0%
0%
5%
0%
0%
0%
0%
N/A
0%
0%
0%
0%
0%
CO
80-120%
n/a
99%
99%
98%
99%
94%
96%
101%
98%
102%
98%
98%
96%
101%
99%
101%
101%
101%
98%
95%
100%
101%
n/a
S02
80-
120%
n/a
83%
70%
80%
65%
65%
68%
65%
66%
63%
71%
78%
77%
79%
65%
69%
72%
73%
75%
79%
66%
70%
n/a
GPS check
meets +/- 10
m criteria
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 1 of 7
6/1/2012
Cicero Rail Yard Study (CIRYS)
Appendix C: Mobile and Stationary Side-by-Side Sampling
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 2 of 7
6/1/2012
This Appendix provides further information on the side-by-side sampling conducted between the mobile
air monitoring vehicle and the air monitoring station located to the NE of the Cicero Rail Yard. The data
shown are for five mobile monitoring sessions conducted on consecutive days, from November 17-21,
2010. For each day, the mobile air monitoring vehicle was parked for 1-2 hours adjacent to the sampling
station. The vehicle sampling height was approximately 1.5 m, whereas the station sampling height was
approximately 2 m. The location of the intercomparison sampling is shown in Figure C-l and is
approximately 50 m from the nearest set of train tracks.
Figure C-l. Intercomparison sampling location.
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 3 of 7
6/1/2012
Sampling date: November 17, 2010
2.5
1.5
x10
O
CQ
-Mobile car Stationary
13:15
13:30
13:45
14:00
14:15
14:30
1000
800
600-
Mobile car - Stationary |
800-
g. 600- I
8 JjsMwo^^
nl I I I I I
13:00 13:15 13:30 13:45 14:00 14:15 14:30
100i i i n—— —r— —r— —|
| Mobile car Stationary-5 min Stationary-hourly
o» 50
^^r
0
13:
00
13:15
13:30
13:45
14:00
14:15
14:30
Figure C-2. Intercomparison time series for BC, CO, and PM2.5 on November 17, 2010.
5000
4000
"5 3000
D)
_£_
2000
CQ
1000
0
- Mobile car Mobile car - 5 min average Stationary
i/Jlk,v,>
>.u
13:00
13:15
13:30
13:45
14:00
14:15
Figure C-3. Intercomparison time series for BC on November 17, 2010, with the y-axis decreased
to 5000 ng m~3 (5 ng m~3).
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 4 of 7
6/1/2012
Sampling date: November 18, 2010
Note: No BC data was available at the stationary monitoring site for this period of time, therefore no
second figure showing the mobile BC data averaged at a 5 minute rate for comparison with the
stationary monitoring data set is provided on 11/18.
jc1Q
3000
2000
O 1000
?:00
07:30
08:00
08:30
Mobile car Stationary
j^^A^A^-w^/y^s^^
7:00
07:30
08:00
08:30
rT 40
'E
§20
" — s
CL 0
_ | - Mobile car Stationary-5 min Stationary-hourly |
07:00
07:30
08:00
08:30
Figure C-4. Intercomparison time series for BC, CO, and PM2.5 on November 18, 2010.
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 5 of 7
6/1/2012
Sampling date: November 19, 2010
6
'E 4
0)
§ 2
CO
1000
Q.
S 500
0
o
09
80
'"T" 60
0) 40
^ 20
CL °
f
x104
| Mobile car Stationary |
J
10 07:20 07:30 07:40 07:50 08
i i i i
| Mobile car Stationary
«. JV~» JWV 1 n > »/\>f(~**, A. ft ^«-il - I 1 L.. Jl AfA
^^^~'"»"W' '"S^J» .-»y»V<'*»-V^— ' " ^~^ --* ~ »V"" -"
00 08
A.'
^_ 7=^^
-
00 08
10
10
10
Figure C-5. Intercomparison time series for BC, CO, and PM2.5 on November 19, 2010.
07:00
07:15
07:30
07:45
08:00
08:15
Figure C-6. Intercomparison time series for BC on November 19, 2010, with the y-axis decreased
to 5000 ng m"3 (5 ng m"3).
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 6 of 7
6/1/2012
Sampling date: November, 20, 2010
6000
o
'E 4000
0)
o
CQ
2000-
08:15
08:30
08:45
09:00
1000
O
o
500
60
40
-Mobile car Stationary
08:15
08:30
08:45
09:00
D.
20
0
-20
| Mobile car Slallonary-5 mln Stationary-hourly j
08:15
08:30
08:45
09:00
09:15
09:15
09:15
09:30
09:30
09:30
Figure C-7. Intercomparison time series for BC, CO, and PM2.5 on November 20, 2010.
5000
4000
"g 3000
05
^
O 2000
OQ
1000
- Mobile car Mobile car - 5 min average Stationary |
i:00
08:15
08:30
08:45
09:00
09:15
Figure C-8. Intercomparison time series for BC on November 20, 2010, with the y-axis decreased
to 5000 ng m~3 (5 ng m~3).
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix C
Page 7 of 7
6/1/2012
Sampling date: November, 21, 2010
6
pO
'£ 4
O)
c
0 2
CO
221
1000
S^
Q_
^ 500
O
O
22s
«~"100
E
jf 50
^
Q- 0
22
x104
I
Mobile car Stationary!
: . I
I
30 22:45 23:00
i
Mobile car Stationary
0 0 0 0 0 I
o o^ ;jl
I
• I
, |
30 22:45 23:00
Mobile car Stationary- 5 mm
_*,
i
. _ ; A U,
23:15 23:30 23:45
I • •
t /, ;
i
^ ° ^ Jl ° A IL! ° _^
i
23:15 23:30 23:45
Stationary-hourly
\
30 22:45 23:00
i
i
23:15 23:30 23:45
Figure C-9. Intercomparison time series for BC, CO, and PM2.5 on November 21, 2010.
5000
4000
'c 3000
O
CO
2000-
1000
22:20
IT
• Mobile car Mobile car - 5 min average Stationary
22:30
22:40
22:50
23:00
23:10
Figure C-8. Intercomparison time series for BC on November 21, 2010, with the y-axis decreased
to 5000 ng m~3 (5 ng m~3).
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 1 of 32
6/25/2012
Cicero Rail Yard Study (CIRYS)
Appendix D: Mobile Monitoring Wind Roses and Driving Maps
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 2 of 32
6/25/2012
Description: Wind roses are provided during each mobile sampling session - one rose for the period
when the vehicle was in driving mode and a second wind rose for the period when the mobile vehicle
was parked at the stationary monitoring location NE of the rail yard. Concentration maps represent
statistically significant excess concentrations in neighborhoods NT1-NT4 in comparison to the urban
background. Concentrations shown are net values - (average value in a specific 50 m increment minus
background mean).
Mobile
session
date:
Wind rose during driving period
Wind rose during stationary period
10/27
Timespan: 10/27/2010 09:16:00 to 10/27/2010 13:05:00
Timespan: 10/27/2010 13:05:00 to 10/27/2010 13:25:00
10/28
Timespan: 10/28/2010 18:53:00 to 10/29/2010 00:50:00
Timespan: 10/28/2010 18:53:00 to 10/29/2010 00:50:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 3 of 32
6/25/2012
10/29
Timespan: 10/29/2010 18:45:00 to 10/29/201022:45:00
Timespan: 10/29/2010 22:45:00 to 10/29/2010 23:45:00
10/30
Timespan: 10/30/2010 08:52:00 to 10/30/2010 12:15:00
Timespan: 10/30/2010 12:15:00 to 10/30/2010 13:15:00
10/31
Timespan: 10/31/2010 03:52:00 to 10/31/2010 07:23:00
Timespan: 10/31/2010 07:25:00 to 10/31/2010 08:25:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 4 of 32
6/25/2012
11/01
Timespan: 11/1/2010 19:18:00to 11/1/2010 23:22:00
Timespan: 11/1/2010 23:22:00 to 11/2/201000:10:00
11/03
Timespan: 11/3/2010 11:50:00 to 11/3/2010 15:25:00
Timespan: 11/3/2010 15:25:00 to 11/3/2010 16:25:00
11/04
Timespan: 11/4/2010 04:10:00 to 11/4/2010 07:45:00
Timespan: 11/4/2010 07:45:00 to 11/4/2010 08:42:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 5 of 32
6/25/2012
11/05
Timespan: 11/5/2010 09:00:00 to 11/5/2010 12:23:00
Timespan: 11/5/2010 13:50:00 to 11/5/2010 14:30:00
11/06
Timespan: 11/6/2010 03:52:00 to 11/6/2010 07:15:00
Timespan: 11/6/2010 07:15:00 to 11/6/2010 08:15:00
11/07
Timespan: 11/7/2010 19:40:00 to 11/7/2010 23:09:00
Timespan: 11/7/2010 23:09:00 to 11/7/2010 23:56:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 6 of 32
6/25/2012
11/08
Timespan: 11/8/2010 19:00:00 to 11/8/2010 22:52:00
Timespan: 11/8/2010 22:52:00 to 11/9/2010 00:10:00
11/10
Timespan: 11/10/2010 09:10:00 to 11/10/2010 12:30:00
Timespan: 11/10/2010 12:30:00 to 11/10/2010 14:00:00
11/11
Timespan: 11/11/2010 04:00:00 to 11/11/2010 07:10:00
Timespan: 11/11/2010 07:10:00to 11/11/2010 09:40:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 7 of 32
6/25/2012
11/12
Timespan: 11/12/2010 10:00:00 to 11/12/2010 13:41:00
Timespan: 11/12/2010 13:41:00 to 11/12/2010 15:05:00
11/13
Timespan: 11/13/2010 04:00:00 to 11/13/2010 07:00:00
Timespan: 11/13/2010 07:00:00 to 11/13/2010 08:40:00
11/15
Timespan: 11/15/2010 19:30:00 to 11/15/2010 23:10:00
Timespan: 11/15/2010 23:10:00 to 11/16/2010 01:05:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 8 of 32
6/25/2012
11/16
Timespan: 11/16/2010 18:55:00 to 11/16/2010 23:30:00
Timespan: 11/16/2010 23:30:00 to 11/17/2010 01:30:00
11/17
Timespan: 11/17/2010 09:45:00 to 11/17/2010 12:52:00
Timespan: 11/17/2010 12:52:00 to 11/17/2010 14:37:00
11/18
Timespan: 11/18/2010 03:58:00 to 11/18/2010 07:00:00
Timespan: 11/18/2010 07:00:00 to 11/18/2010 08:42:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 9 of 32
6/25/2012
11/19
Timespan: 11/19/2010 03:52:00 to 11/19/2010 07:08:00
Timespan: 11/19/2010 07:08:00 to 11/19/2010 08:30:00
11/20
Timespan: 11/20/2010 03:57:00to 11/20/2010 08:06:00
Timespan: 11/20/2010 08:06:00 to 11/20/2010 09:49:00
11/21
Timespan: 11/21/2010 19:02:00 to 11/21/2010 22:30:00
Timespan: 11/21/2010 22:30:00 to 11/22/2010 00:09:00
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 10 of 32
6/25/2012
Sampling date: 10/27
Timespan: 10/27/2010 09:16:00 to 10/27/2010 13:05:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
H70
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m )
i FEch
900
800
700
600
500
400
300
200
100
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultraflne Particles (p cm"3)
n
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
PM10(ngrrfJ
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 11 of 32
6/25/2012
Sampling date: 10/28
Timespan: 10/28/2010 18:53:00 to 10/29/2010 00:50:00
Carbon Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m )
P
800
700
600
500
400
300
200
100
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM10(ngrrf
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785
n
2.6
2.4
2.2
2
1.8
1.6
1.4
1.2
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 12 of 32
6/25/2012
Sampling date: 10/29
Timespan: 10/29/2010 18:45:00 to 10/29/201022:45:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-360
-340
-320
-300
-280
-260
-240
-220
-200
- 180
- 160
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m"3)
f!
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm )
113000
112000
111000
j10000
-9000
-8000
-7000
-6000
-5000
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
PM2,(ngm-J)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
n
PM10(ngrrf-
5
4.5
4
3.5
3
2.5
2
1.5
1
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 13 of 32
6/25/2012
Sampling date: 10/30
Timespan: 10/30/2010 08:52:00 to 10/30/2010 12:15:00
41.854
41.852
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
- 200
- 180
- 160
- 140
- 120
- 100
- 80
- 60
- 40
41.854
41.852
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m )
41.854
41.852
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafne Particles (p cm'3)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
X10
~ 1.2631
I 1.2631
I 1.263
I 1.263
I 1.263
- 1.263
- 1.263
- 1.2629
- 1.2629
- 1.2629
PM10(ngrrT3)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 14 of 32
6/25/2012
Sampling date: 10/31
Timespan: 10/31/2010 03:52:00 to 10/31/2010 07:23:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
N/A - failed quality check
785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM,0(|4gnrfJ)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
- 1
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 15 of 32
6/25/2012
Sampling date: 11/01
Timespan: 11/1/2010 19:18:00to 11/1/2010 23:22:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
o.8
r
•J0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m
1000
900
800
700
600
500
400
300
200
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"J)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
- 1.1
- 1.05
- 1
-0.95
-0.9
-0.85
-0.8
^25^3 m"3'
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
n
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
- 1
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 16 of 32
6/25/2012
Sampling date: 11/03
Timespan: 11/3/2010 11:50:00 to 11/3/2010 15:25:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-320
-300
-280
-260
-240
-220
-200
- 180
- 160
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m"3)
fl
2000
1800
1600
1400
1200
1000
800
600
400
200
-87.78 -87.77 -87.76 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm )
18500
18000
17500
-pooo
Jesoo
Jeooo
-5500
-5000
-4500
-4000
-87.78 -87.77 -87.76 -87.75 -87.74
PM25(ngm~-
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
n
-87.79 -87.78 -87.77 -87.76 -87.75
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
Jo.9
Jo.8
L
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 17 of 32
6/25/2012
Sampling Date: 11/04
Timespan: 11/4/2010 04:10:00 to 11/4/2010 07:45:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75
•J0.9
Jo.8
1°'7
Jo.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m )
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm
i 13000
12000
11000
10000
- 9000
- 8000
- 7000
- 6000
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 18 of 32
6/25/2012
Timespan: 11/5/2010 09:00:00 to 11/5/2010 12:23:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
394
393.8
393.6
393.4
393.2
393
392.8
392.6
392.4
392.2
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
Black Carton (ng m"3)
Ultrafne Particles (p cm"J)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM25(ngm'J)
41.844
41.842
41.84
41.838
41.836
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 19 of 32
6/25/2012
Sampling date: 11/06
Timespan: 11/6/2010 03:52:00 to 11/6/2010 07:15:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
Black Carbon (ng m"3)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM25(ngm'J)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 20 of 32
6/25/2012
Sampling date: 11/07
Timespan: 11/7/2010 19:40:00 to 11/7/2010 23:09:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
755.8
- 755.6
- 755.4
- 755.2
- 755
- 754.8
- 754.6
- 754.4
- 754.2
- 754
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM10(ngrrf3)
n
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 21 of 32
6/25/2012
Sampling Date: 11/08
Timespan: 11/8/2010 19:00:00 to 11/8/2010 22:52:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
3SO
-340
-330
-320
-310
-300
-290
-280
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"3)
18500
18400
18300
18200
-UlOO
-8000
-7900
-7800
-87.79 -87.78 -87.77 -87.76 -87.75
PM25(ngm'J)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78
-87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
- 1
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 22 of 32
6/25/2012
Sampling date: 11/10
Timespan: 11/10/2010 09:10:00 to 11/10/2010 12:30:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m"*
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"1
fl
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
X10
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
PM,,(ngrrr3)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 23 of 32
6/25/2012
Sampling date: 11/11
Timespan: 11/11/2010 04:00:00 to 11/11/2010 07:10:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
\0.9
Jo.8
I0'7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m"3)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm'3)
-87.79
-87.78 -87.77 -87.76 -87.75 -87.74
110500
110000
19500
-Uooo
-Usoo
- 8000
- 7500
- 7000
- 6500
- 6000
- 5500
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
- 4.2
- 4
- 3.8
- 3.6
- 3.4
- 3.2
- 3
- 2.8
- 2.6
; 2.4
-87.79 -87.78 -87.77 -87.76
-87.75 -87.74
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 24 of 32
6/25/2012
Sampling date: 11/12
Timespan: 11/12/2010 10:00:00 to 11/12/2010 13:41:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75
- 180
- 170
- 160
- 150
- 140
- 130
- 120
- 110
- 100
-90
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m"3)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafne Particles (p cm"3)
1.2
1
-0.8
-0.6
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
PM2 5 (ng rrf-
PM10(ngrrfJ
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
2.8
2.6
2.4
2.2
2
1.8
1.6
1.4
1.2
1
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75
I 0.9
0.8
07
0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 25 of 32
6/25/2012
Sampling date: 11/13
Timespan: 11/13/2010 04:00:00 to 11/13/2010 07:00:00
Carbon Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"3)
fl
13000
12000
11000
10000
9000
8000
7000
6000
5000
4000
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 26 of 32
6/25/2012
Sampling date: 11/15
Timespan: 11/15/2010 19:30:00 to 11/15/2010 23:10:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m )
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"3)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
x 10
! 1.741
1.741
1.7409
1.7409
1.7409
1.7409
1 1.7409
- 1.7408
- 1.7408
: 1.7408
PM10(ngm-3)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
1.26
1.24
1.22
1.2
1.18
1.16
1.14
1.12
1.1
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 27 of 32
6/25/2012
Sampling date: 11/16
Timespan: 11/16/2010 18:55:00 to 11/16/2010 23:30:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75
1190.2
190
189.8
189.6
189.4
- 189.2
- 189
- 188.8
- 188.6
- 188.4
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m
800
750
700
650
600
550
500
450
400
350
300
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm'J)
-87.79 -87.78 -87.77 -87.76 -87.75
PM10(ngm-3)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
3.6
3.4
3.2
3
2.8
2.6
2.4
2.2
2
1.8
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75
Jo.9
Jo.8
r
Jo.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 28 of 32
6/25/2012
Sampling date: 11/17
Timespan: 11/17/2010 09:45:00 to 11/17/2010 12:52:00
Carton Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m )
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM10(ngm-J)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
- 1
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
-0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 29 of 32
6/25/2012
Sampling date: 11/18
Timespan: 11/18/2010 03:58:00 to 11/18/2010 07:00:00
Carton Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834'
-87.79
Black Carbon (ng m~*
-87.78 -87.77 -87.76 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.785
-H.8
J1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 30 of 32
6/25/2012
Sampling date: 11/19
Timespan: 11/19/2010 03:52:00 to 11/19/2010 07:08:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carbon Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1132.8
132.6
132.4
132.2
- 132
- 131.8
- 131.6
- 131.4
- 131.2
• 131
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng rrf"
900
800
700
600
500
400
300
200
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"3)
x 10
- 1.9
- 1.8
- 1.7
- 1.6
- 1.5
- 1.4
- 1.3
- 1.2
- 1.1
- 1
-0.9
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
PM25(ngrrfJ)
PM10(ngrrfJ)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
I0.9
0.8
07
0.6
- 0.5
- 0.4
- 0.3
- 0.2
- 0.1
- 0
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 31 of 32
6/25/2012
Sampling date: 11/20
Timespan: 11/20/2010 03:57:00to 11/20/2010 08:06:00
Carbon Monoxide (ppb)
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carbon (ng m"*
244
242
240
238
236
234
232
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
PM25(ngrrrJ)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-87.785 -87.78 -87.775 -87.77 -87.765 -87.76 -87.755 -87.75
-------
Cicero Rail Yard Study (CIRYS) Report - Appendix D
Page 32 of 32
6/25/2012
Sampling date: 11/21
Timespan: 11/21/2010 19:02:00 to 11/21/2010 22:30:00
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Carton Monoxide (ppb)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Black Carton (ng m
550
500
450
400
350
300
250
200
150
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
Ultrafine Particles (p cm"3)
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
x 10
11.5
"
-0.9
-0.8
-0.7
PM25(ngm-3)
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
41.85
41.848
41.846
41.844
41.842
41.84
41.838
41.836
41.834
F
- 1.8
- 1.6
- 1.4
- 1.2
- 1
-87.79 -87.78 -87.77 -87.76 -87.75 -87.74
------- |