suzanne paulson* & nico schulte** · identifying urban designs and traffic management...
TRANSCRIPT
IDENTIFYING URBAN DESIGNS AND TRAFFICMANAGEMENT STRATEGIES FOR SOUTHERN CALIFORNIA
THAT REDUCE AIR POLLUTION EXPOSURE
Suzanne Paulson* & Nico Schulte**Wonsik Choi,* J.R. DeShazo,* Dilhara Ranasinghe,*
Akula Venkatram,** Lisa Wong,* Owen Hearey,* Karen Bunavage,* Rodrigo Siguel,* Arthur Winer,*
Mario Gerla,* Kathleen Kozawa,*** Steve Mara,*** Si Tan**
*UCLA Departments of Atmospheric and Oceanic Sciences, Institute of Environment and Sustainability,
Public Policy, Luskin Center for Innovation, Computer Science & Environmental Health Sciences
**UCR Department of Mechanical Engineering
***CARB
1
Objectives• Develop guidelines for TOD planners to reduce pedestrian
exposure to air pollution in urban built environments through field measurements, analysis and statistical modeling.
• Develop a dispersion model that can be used to provide TOD planners quantitative links among the variables that control dispersion in complex urban environments.
2
Air pollutants from traffic sources
4
• Oxides of Nitrogen• Volatile Organic Compounds, including some air toxics• Ultrafine Particles (UFP)• Brake dust• UFP and the gasses undergo similar dispersion in the
atmosphere.
Pollutant Concentrations Near Roadways Vary A LOT• Fleet emissions
• Traffic density, fleet composition, driving conditions• Atmospheric Dispersion
• Wind speed and direction, atmospheric stability, vehicle wakes, topography
• Built Environment• Roadway geometry, buildings, soundwalls, other
nearby roadways & vegetation• Observed roadway pollutant spatial distributions also
depend on the relationship between the peak and background concentrations.
5
Fine Coarse
Very fine dust from mechanical processes
Mostly formed in the atmosphere
Directly emitted or formed in the atmosphere
Very Small Particles: lots of spatial heterogeneity
Ultrafine particles are mostly from vehicular emissions. They disappear In around a half an hour: rather than magically going away, they collide and stick to fine particles. As a result, they are highly elevated around roadways compared to everywhere else.
Plot Source: Wilson et al. (1977)
Ultrafine
“Par
ticle
Mas
s”
Particle Diameter (microns)
6
Tropospheric Aerosols are Complex Mixtures
Numbers range from 103 to 107 cm-1 in urban areas
Soot Agglomerate
Liquid Droplet
Included Accumulation Mode Particles Diesel
Exhaust
Diesel Particle Close-up
PM2.5 particle
7
Oberdorster, Oberdorster & Oberdorster 2005; Int’l Commission on Radiological Protection 1994
τ = 1 yr for 50% clearance
Where Particles Deposit Depends on Their Size8
1µm0.1µm(and smaller)
E.R. Weibel, University of Bern
TRANSLOCATION FROM AIR TO BLOOD
Courtesy of Peter Gehr, U. Bern
9
Multiple Pathways to Increased Morbidity in Children are Associated with Proximity to Traffic
Prenatal Impacts Los Angles - women living near high heavy duty traffic areas were
at increased risk of premature delivery and low birth weight babies (Ritz and Co-workers, UCLA)
Asthma Prevalence Southern California - prevalence of asthma among children was
associated with several indicators of exposure to traffic including proximity of the home to a freeway
Respiratory Symptoms East Bay (San Francisco area) - children attending schools near
freeways had more respiratory symptoms
10
Multiple Pathways to Increased Morbidity & Mortality in Adults are Associated with Proximity to Traffic
• Acute respiratory diseases; • acute asthma, chronic obstructive pulmonary disease, pneumonia,
lung cancer
• Cardiovascular disease; • Heart attacks and stroke
• Many other diseases • Air pollution degrades overall health, beginning prior to birth. This
results in higher incidences of many diseases and conditions.
11
What makes UFPs distinct from other pollutants includes:
Source strengtha wide variability in particle number emission factors which depend on fuel and engine types, maintenance and age of vehicles, driving conditions, etc
[Morawska et al., 2008]
Dynamic nature
short lifetime (~ hours) and on-going transformation of their physical and chemical properties including number and size distributions which decreases with increasing particle size
[Birmili et al., 2013; Choi et al., 2016a]
Atmos. dispersion
governed by wind speed/direction, atmospheric stability, and “breathability” of cities which depends on the urban morphology (built-environment)
[Buccolieri et al., 2010; Choi et al., 2016b]
Substantial temporal and spatial variability of UFPs distributions both in number and sizes
Spatial heterogeneity of UFPs (and inventory) remains a key challengefor the assessment of their impacts on human health and climate
12
Instrument Measurement ParameterCPC (TSI, Model 3007) UFP number concentration (10 nm−
1µm)FMPS (TSI, Model 3091) Particle size distribution (5.6−560 nm)DustTrak (TSI, Model 8520)
PM2.5 and PM10 mass
EcoChem PAS 2000 Particle bound PAHsLI-COR, Model LI-820 CO2
Teledyne API Model 300E
CO
Teledyne API Model 200E
NOX
Teledyne API Model 400A
O3
3D-Sonic Anemometer(Campbell CSAT3)
Temperature, Relative humidity, Wind speed/direction, Turbulence Characteristics
Garmin GPSMAP 76CS GPSSmartTetherTM Vertical profiles of temperature, RH,
wind speed/directionKciVacs video Video record for traffic and fleet
composition
California Air Resources Board Mobile
Measurement platform(MMP)
Toyota RAV4 electric vehicle
Mobile Monitoring Platform
• Six field sites in LA county 2013 – 2014
• Parameters measured:• Meteorological
measurements (3D sonic)• Near road ultrafine particle
number concentration measurements (CPC)
16
Stationary Measurements
“Roadmap”• Introduction• The impact of the dimensions on pollution in street
canyons (N. Schulte)• Results of high spatial resolution (~5 m) mobile &
stationary measurements made in 5 neighborhoods in Los Angeles• High fidelity processing of mobile measurements• A microdynamics model for predicting concentrations along the
street• The neighborhood view: The impact of the design of the built
environment on the scale of several large city blocks• Benefits of locating the bus stops a bit further from the
intersections• Summary for Planners
17
Street Canyons
Nico Schulte, Si Tan and Akula Venkatram (2015) The ratio of effective building height to street width governs dispersion of local vehicle
emissions. Atmos. Environ. 11:54–63.
19
Objective and Model ApplicationTo develop models that account for building effects in linking vehicle emissions to near road pollutant concentrations.
Models validated with field data can be used to conduct numerical experiments that would be impossible in the real world.
Evaluate impact of designs for transit oriented development on near-road concentrations of traffic emissions
Evaluate mitigation methods including traffic management, limiting building height, creating open space, pedestrian zones
21
Approach
22
Formulate model based on past
research
Design field study using model to guide required
measurements
Conduct field study to collect
dataEvaluate model
with data
Modify model using results from evaluation
Vertical Dispersion Model
23
Q Street emission rate
Cs Surface concentration averaged over the street
Cr Roof concentrationW Street widthH Building heightar Aspect Ratio (H/W)σw Average standard
deviation of vertical velocity fluctuations
β Empirical constanth0 Initial vertical
mixing
0
1
1 (1 )w
Q rs rW
r
aC Ch aH
βσ
+
= + + +
Area Weighted Building HeightL Length of streethi Height of building ibi Length of building i along street
24
1i iL
iH h b= ∑
Left: Google earth view of 8th St LA field site. Right: Building heights along 8th St
Evaluation with LA county data
• Six field sites in LA county 2013 –2014
• Near road ultrafine particle number concentration measurements
25
Left: Scatter plot (normalized by emission rate)Right: q-q plot. 30 minute average UFP concentrations
Evaluation with Riverside data
• August/September 2015 Riverside Market Street
• Carbon monoxide and UFP measurements
26
Top: Scatter plot, Bottom: q-q plot. 2 hour average CO concentrations
Evaluation with Riverside data
27
2 hour average CO concentrations normalized by traffic emission rate
Estimation of VDM Inputs
• How can VDM model inputs be estimated?• Meteorology not routinely measured in urban area.• Develop model to relate routine meteorological
measurements at an upwind reference location with the necessary urban values.
28
Estimation of VDM Inputs
z Height from groundu* Surface friction velocityL Monin-Obukhov lengthz0 Surface roughness lengthU Wind speedψm Stability modification to wind
speed profile
• Internal boundary layer model describes evolution of turbulence from rural to urban area.
29
(Fisher et al. 2006)𝑑𝑑𝑑𝑑𝑑𝑑𝑑
= 𝐴𝐴 𝜎𝜎𝑤𝑤𝑈𝑈𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢(𝑑)
(Garratt 1990)
𝑈𝑈𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 ℎ = 𝑈𝑈𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑟𝑟(ℎ)=> 𝑢𝑢∗𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢
𝑈𝑈 𝑧𝑧 =𝑢𝑢∗𝜅𝜅
ln𝑧𝑧𝑧𝑧0
+ 𝜓𝜓𝑚𝑚𝑧𝑧0𝐿𝐿
− 𝜓𝜓𝑚𝑚𝑧𝑧𝐿𝐿
Implications for Policy
• Ratio of area weighted building height to street width, the effective aspect ratio, governs near-road concentrations.
• VDM suggests that the effective aspect ratio of streets with high local vehicle traffic should be limited to reduce exposure to elevated concentrations of traffic emissions.
30
Project Outcome – UCR VDM Tool• Combine VDM + AERMOD for
practical application by planners
31
Develop project level emission estimates
Use AERMOD to compute Cr (with standard urban
options)
Compute building morphological parameters
(from digital elevation model)
Use VDM tool to compute surface concentration CsVDM GUI interface
Model is available to download at https://www.arb.ca.gov/research/single-project.php?row_id=65135
Processing Mobile Data
Ranasinghe, D., W.S. Choi, A.M. Winer and S.E. Paulson (2016) Developing High Spatial Resolution Concentration Maps Using Mobile
Air Quality Measurements. Aerosol and Air Qual. Res. 16 (8), 1841-1853.
Mobile Data Challenges• Mobile data gives spatially heterogeneous measurements; sometimes you get 30 measurements in one place; sometimes one every 20 m.
• Simple averaging of mobile data (after correction for the wandering GPS signal) ends up looking like a trail of confetti after a parade route.
• How many runs do you need?
Using a line-reference system• Divide the street into a grid with reference points every x
meters. • Each reference point gets 1 value per run. If there are 30 data
points, we average them. If there are no data points, we interpolate a data point.
• This avoids under/overweighting individual “runs” on the route.
MMProuterun 1run 2run 3run 4
Need ~20 repeats under similar met conditions to get a reasonable average
Ranasinghe et al. AAQR (2016)
Morning
With high emitter spikes High emitter spikes removed
Afternoon
Some features appear consistently
Modeling the Determinants of Highly-localized UFP
Concentrations
J.R. DeShazo, Suzanne Paulson, Lisa Wu, Owen Hearey, and others
39
Data uses in Model• April-July 2008 • 2nd St. to Jefferson Blvd.• ~3 miles long • 12 MMP runs • ~7000 observations
Broadway Transect in Downtown Los Angeles
40
Why take this approach?1. How much of the observed variation UFP can be
attributed to each explanatory variable? How accurately can we measure these relationships?
2. How much of the observed UFP can be explained by these localized factors? Which local factors we leaving out? How much of the observed variation cannot be explained by accounting for local factors?
3. Estimated model can be used to prediction/simulate concentrations at any location for any configuration and level of factors.
41
Modeling the effects of highly-localized factors UFP = f(emission sources,
+ atmospheric conditions, + built environment, + position of the MMP, + constant (background concentration)
Regression model will seek to explain the concentration of UPF measure as function of a these explanatory variables.
43
Validating the model’s predictive power: Actual v. Predicted UPF Concentrations for AM and PM Transects
a. b.
44
Three types of findings• 1. Cumulative impacts of factors on street level UFP,
• 2. Dynamic UFP patterns associated with common traffic events,
• 3. Predicted free-flow versus stop start UFP spatial and temporal dynamics
45
1. Cumulative impacts of factors on street level UFP
• How much does each of these factors explain that variation observed in UFP?
• We estimate the cumulative effect of a given variable on UFP by placing all of the other parameters at their mean values. We then estimate the predicted concentration of UFP when the variable of interest is set to zero, and compared that with predicted UFP concentration when the variable of interest is set to its mean level.
• Next slide highlights the top five factors
46
Table 4: Cumulative Impacts of Traffic EventsImpact Std. Error 95% Conf. Interval
MMP State of Motion and SpeedIdle-to-moving -28.65 (-11.4) -50.99 -6.31Speed (m/s) -15.32 (-5.74) -26.56 -4.07Lane number 15.73 (-4.47) 6.97 24.49
On-going TrafficLight-duty 9.78 (-14.01) -17.69 37.24Medium & Heavy-duty 142.76 (-90.31) -34.26 319.78Buses -25.66 (-155.91) -331.25 279.92Acceleration event -58.7 (-122.98) -299.74 182.34
On-Coming TrafficLight-duty -0.99 (-16.77) -33.85 31.87Medium & heavy-duty -18.8 (-57.18) -130.88 93.28Buses -21.31 (-71.44) -161.34 118.72Acceleration event -158.72 (-127.01) -407.66 90.22
Crossing from the Left TrafficLight-duty 30.05 (-34.74) -38.04 98.15Medium & heavy-duty 80.79 (-76.07) -68.31 229.89Buses 92.64 (-111.96) -126.8 312.08Acceleration event 26.96 (-98.47) -166.05 219.97
Crossing from the Right TrafficLight-duty -10.94 (-58.6) -125.8 103.92Medium & heavy-duty -30.58 (-162.81) -349.68 288.53Buses 54.04 (-59.01) -61.61 169.69Acceleration event 52.59 (-50.29) -45.98 151.17
Built EnvironmentIntersection 320.15 (-120.69) 83.6 556.7Average building height (m) 1.7 (-4.5) -7.13 10.52Building height differential (m)a 233.06 (-270.62) -297.36 763.48
47
2. Dynamic UFP patterns associated with common traffic events • Model Characterizes Dynamic Traffic Events:
• UFP patterns associated with • Traffic crossing at an intersection in front of the MMP as it is
stopped at a light,• Oncoming heavy duty vehicles in traffic while the MMP travels• Ongoing and oncoming accelerations events from the stopped
position while the MMP accelerates from a stopped position traveling with ongoing traffic.
48
We do two planning/policy simulations
A. What happens to free-flow vs stop/start?
B. What happens when there are more or less light duty vehicles, heavy duty, etc
3. Predicted free-flow versus stop start UFP spatial and temporal dynamics
51
• MMP cruises for 225m/30 sec. at constant speed (7.5 m/s) across an intersection
• All neighboring buildings are 15m high; the intersection is 12m across (from 103 to 115)
• Wind and location fixed effects are set to long-run averages
• No acceleration instances occur, and no cross-traffic is encountered
• Ongoing and oncoming traffic are set to long-run averages
“Free-flow” simulation
52
• MMP travels 225m -- cruises for 10 sec., decelerates for 10 sec., stops for 30 sec., accelerates through the intersection for 10 sec., and cruises for 10 sec.
• As soon as the MMP comes to a complete stop, cross-traffic accelerates into the intersection; as soon as the MMP starts accelerating, ongoing and oncoming traffic accelerate into the intersection
• Oncoming, ongoing, left- and right-cross-traffic (when occuring) are set to long-run averages
“Stop-and-start” simulation
54
Choi, W., D. Ranasinghe, K. Bunavage, J.R. DeShazo, L. Wu, R. Seguel, A.M. Winer, and S.E. Paulson (2016)
The effects of the built environment, traffic patterns, and micrometeorology on street level ultrafine particle concentrations
at a block scale: Results from multiple urban sites. Sci. Tot. Environ. 15;553:474-85. doi: 10.1016/j.scitotenv.2016.02.083.
What is the effect of the built environment at the
block/neighborhood scale on pollutant concentrations at the
street?
57
Roof-top Sonic tower
SurfaceSonic tower
Mid Block Mid Block
Mid Block Intersection
Intersection
DiSCminiMMP route
Video-cameraMeasurement Design 58
-118.26 -118.255 -118.25 -118.245
34.038
34.04
34.042
34.044
34.046
34.048
34.05
34.052
Longitude ( o )
Latit
ude
( o )
0
100
200
300
400
500
Site 1: Street Canyon
Broadway & 7th Site (Street view: heading South)
Building height(Ft.)
59
-118.265 -118.26 -118.255
34.034
34.036
34.038
34.04
34.042
34.044
Longitude ( o )
Latit
ude
( o )
0
50
100
150
200
250
300
350
Olive & 12th Site (Street view: heading to North)
Site 2: One isolated tall building with low trafficBuilding height (Ft.)
60
-118.296 -118.292 -118.28834.056
34.057
34.058
34.059
34.06
34.061
34.062
34.063
34.064
Longitude ( o )
Latit
ude
( o )
50
100
150
200
250
300
350
Vermont & 7th Site (Street view: heading to West)
Site 3: One isolated tall building with high traffic
Building height(Ft.)
61
-118.284 -118.282 -118.28 -118.278 -118.27634.056
34.058
34.06
34.062
34.064
Longitude ( o )
Latit
ude
( o )
0
50
100
150
200
Wilshire & Carondelet Site (Street view: heading to East)
Site 4: Intermediate buildings in one side and low buildings in the other side of the street
Building height(Ft.)
62
-118.0662 -118.0622 -118.058234.1021
34.1041
34.1061
34.1081
34.1101
Longitude ( o )
Latit
ude
( o )
10
15
20
25
30
35
Temple City & Las Tunas Site (Street view: heading to North)
Site 5: All single story buildings
Building height(Ft.)
63
Built environment quantitative descriptors Broadway
& 7th
(Site1)
Olive St. &
12th St. (Site2)
Vermont &
7th St. (Site3)
Wilshire &
Carondelet (Site4)
Temple City &
Las Tunas (Site5)
# of buildings 59 34 90 44 143
Max. building height (m) 58 129 80 57 8
Mean building height, Hbldg (m)
34 21 11 18 5
Bldg area weighted height, Harea (m)
40 42 25 24 6
Bldg. homogeneity, Harea/Hbldg (dimensionless) (1=perfectly homogeneous)
1.16 2.01 2.21 1.39 1.09
Mean building ground area (m2)
1,030 1,395 585 992 225
Street width (m) 26 (BW) / 22 (7th)
28 (Olive) / 17 (12th)
30 (Ver) / 25 (7th)
17 (Car) / 37 (Wil)
24 (TC) / 30 (LT)
Simple Aspect ratio (Harea/Wstreet)
1.7 1.9 0.9 0.9 0.2
Block length (m) 190 (BW) / 100 (7th)
180 (Olive)/ 95 (12th)
190 (Ver) / 95 (7th)
160 (Car) / 75 (Wil)
175 (TC) / 115 (LT)
Ratio occupied by bldg. 0.72 0.42 0.33 0.46 0.30
64
(a) Morning (b) Afternoon
Higher traffic generally higher UFP. In the morning howeverthere are deviations esp. for the two sites with extreme built-environments: the street canyon (Site1) and the low, flat bldg. canopy (Site 5).
65
0 10 20 30 40 50 60 7010
15
20
25
30
35
40
45
[UFP
] (
×10
3pa
rtic
les
⋅cm
-3)
Traffic flow rate (vehicles⋅min-1)10 20 30 40 50 60 70
10
15
20
25
30
35
40
45
[UFP
] (
× 103
part
icle
s⋅c
m-3
)
Traffic flow rate (vehicles⋅min-1)
Site 1Site 2Site 3Site 4Site 5
(a) (b)
Best Explanatory Factor in the Morning: The “Areal Aspect Ratio” =
Length scale of buildings over length scale of open space
( ) ( ) open
bldg
siteopendiag
bldg
sitebldgdiag
bldgarea L
HAAL
HAFL
HAr =
×=
−×=
∑ //1
Hbldg: Mean area-weighted building height
Ldiag: Diagonal length of block
Fbldg: Footprint of the Buildings
Asite: Area of the sampling site
Aopen: Area of the open space in sampling site
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80
200
400
600
800
1000
1200
1400
Areal aspect ratio (Ararea )
[UFP
] (pa
rtic
les ⋅
cm-3
) /
Traf
fic fl
ow ra
te (v
eh. ⋅m
in-1
)
Site 1Site 2Site 3Site 4Site 5
Choi et al., 2016
66
Best Explanatory Factor in the Afternoon: Turbulence strength (vertical fluctuations of surface winds, σw)
1800
n-1)
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.85
10
15
20
25
30
35
40
σw (m⋅s-1)
[UFP
] (×
103 p
artic
les ⋅
cm-3
)
Site 1Site 2
Site 3Site 4Site 5
(a)
67
Best Explanatory Factor in the Afternoon: Turbulence strength (vertical fluctuations of surface winds, σw)
Appears to be from non-local emissions
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80
200
400
600
800
1000
1200
1400
1600
1800
σw ( m⋅s-1)
[UFP
] (pa
rtic
les ⋅
cm-3
) /
Traf
fic fl
ow ra
te (
veh.
⋅min
-1)
Fitting curvesSite 1Site 2Site 3Site 4Site 5
σw (m s )
(b)
68
The effects of building heterogeneity on turbulence in the afternoon:
Higher building heterogeneity appears to enhance surface turbulence, under conditions with moderate winds and an unstable atmosphere
1 1.2 1.4 1.6 1.8 2 2.20
0.5
1
1.5
2
2.5
Building heterogeneity (Harea / Hbldg)
Turb
ulen
ce k
inet
ic e
nerg
y (T
KE)
, m
2 ⋅s-2
1 1.2 1.4 1.6 1.8 2 2.20.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Building heterogeneity (Harea / Hbldg)
σw
(m
⋅ s-1
)
Site 1
Site 2
Site 3
Site 4
Site 5
(a) (b)
σ w(m
∙s-1
)
Turb
ulen
ce k
inet
ic e
nerg
y (m
2 ∙s-1
)
69
Decay of pollutants around the intersections: the best
place for the bus stop?Choi, W.S., D. Ranasinghe, J.R. DeShazo, J.J. Kim and S.E.
Paulson (2017) Cross-Intersection Profiles of Ultrafine Particles in Different Built Environments: Implications for Pedestrian Exposure and Bus Transit Stops. Submitted.
70
Variety of Intersections; 1,744 Profiles TotalWilshire in
Beverly Hills
(5 inter-sections)
Broadway & 7th
Downtown Los
Angeles
Olive & 12th
Downtown Los
Angeles
Vermont & 7th
Wilshire & Carondelet
Temple City & Las
Tunas
Street width
30 - 38 m 22 & 26 m 17 & 28 m 25 & 30 m 17 & 37 m 24 & 30 m
Traffic flow rate (A.M.)
24 12 & 15 21 & 4 39 & 10 31 & 31 25 & 28
Traffic flow rate (P.M.)
47 20 &20 8 & 3 38 & 12 2 & 27 26 & 29
Traffic density
Long queues, WB in A.M., EB in P.M.
Medium queues, slow vehicle speeds
Minimal queues
Long queues, often for entire block
Short queues
Long queues but queues dissipate rapidly
Distance between traffic lights
330 m 125 - 200 m
(1) 180 m(2) 125 m
(1) 224 m(2) 174 mc
(1) 190 m(2) 100 m
(1) 200 m(2) 135 m
73
-80 -60 -40 -20 0 20 40 60 800.8
1.2
1.6
2.0
2.4
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
East-boundWest-bound
-80 -60 -40 -20 0 20 40 60 801.0
1.4
1.8
2.2
2.6
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
-80 -60 -40 -20 0 20 40 60 80
2.4
2.8
3.2
3.6
4.0
4.4
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
North-boundSouth-boundEast-boundWest-boundAveraged
-80 -60 -40 -20 0 20 40 60 801.5
2.0
2.5
3.0
3.5
4.0
4.5
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
North-boundSouth-boundEast-boundAveraged
-80 -60 -40 -20 0 20 40 60 80
2.0
2.5
3.0
3.5
Distance from the intersection center (m)[U
FP] (
×10
4 par
ticle
s ⋅ c
m-3
)
-80 -60 -40 -20 0 20 40 60 801.1
1.3
1.5
1.7
1.9[U
FP] (
×10
4 par
ticle
s ⋅ c
m-3
)
Distance from the intersection center (m)
(a) Beverly
(b) Broadway
(c) Olive
N=355
N=92
N=104
N=245
N=76
N=107
Cross-intersection profiles of UFPs for each traffic directionEarly
morningsAfternoons
74
75
Distance from the intersection center (m)
Distance from the intersection center (m)
-80 -60 -40 -20 0 20 40 60 802.0
3.0
4.0
5.0
6.0
7.0
[UFP
] ( ×
104 p
artic
les
⋅ cm
3 )
Distance from the intersection center (m)
North-boundSouth-boundEast-boundWest-boundAveraged
-80 -60 -40 -20 0 20 40 60 801.0
2.0
3.0
4.0
5.0
6.0
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
-80 -60 -40 -20 0 20 40 60 802.2
2.6
3.0
3.4
3.8
[UFP
] ( ×
104 p
artic
les
⋅ cm
3 )
Distance from the intersection center (m)
North-boundSouth-boundEast-boundWest-boundAveraged
-80 -60 -40 -20 0 20 40 60 801.5
2.0
2.5
3.0
3.5
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
-80 -60 -40 -20 0 20 40 60 801.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
[UFP
] ( ×
104 p
artic
les
⋅ cm
3 )
Distance from the intersection center (m)
North-boundSouth-boundEast-boundWest-boundAveraged
-80 -60 -40 -20 0 20 40 60 80
2.0
3.0
4.0
5.0
6.0
7.0
[UFP
] ( ×
104 p
artic
les
⋅ cm
-3)
Distance from the intersection center (m)
(d) Vermont
(e) Wilshire
(f) Temple City
N=79
N=85
N=184
N=143
N=101
N=181
Cross-intersection profiles of UFPs for each traffic direction
76
Distance from the center of Intersection
Distance from the center of Intersection
-80 -60 -40 -20 0 20 40 60 80
[UFP
] (pa
rticl
es c
m-3
)
1.8e+4
1.9e+4
2.0e+4
2.1e+4
2.2e+4
2.3e+4
2.4e+4
2.5e+4
2.6e+4
stan
dard
dev
iatio
n (1
σ)
0
2e+4
4e+4
6e+4
8e+4
1e+5
Distance from the center of Intersection
-80 -60 -40 -20 0 20 40 60 80
[UFP
] (pa
rticl
es c
m-3
)
1.8e+4
1.9e+4
2.0e+4
2.1e+4
2.2e+4
2.3e+4
2.4e+4
2.5e+4
2.6e+4
stan
dard
dev
iatio
n (1
σ)
2e+4
3e+4
4e+4
5e+4
6e+4
7e+4
N=85335.0%
N=174426.5%
Distance from the center of Intersection
Distance from the center of Intersection
-80 -60 -40 -20 0 20 40 60 80
[UFP
] (pa
rtic
les
cm-3
)
1.8e+4
1.9e+4
2.0e+4
2.1e+4
2.2e+4
2.3e+4
2.4e+4
2.5e+4
2.6e+4
stan
dard
dev
iatio
n (1
σ)
2e+4
3e+4
4e+4
5e+4
6e+4
7e+4
N=174426.5%
Average Profiles
0.25 0.5 0.75 1
104
105
Cumulative Distribution
[UFP
] (pa
rtic
les⋅
cm-3
)
[UFP] at peak location[UFP] at base locationData used for a linear fit at peakData used for a linear fit at baseExtended fit at peakExtended fit at base
Cumulative distributions of UFPs at the peak and base locations of the profile
77
0 25 50 75 100 125 150 175 200
% elevation of [UFP] near intersection compared to 40 m away
-20
-10
0
10
20
30
40
50
60
70
% re
duct
ion
of U
FP e
xpos
ure
leve
l (%
)Slow walking (0.5 m s -1
), stop at 40 m
Comfortable walking (1.0 m s -1), stop at 40 m
Normal walking (1.5 m s -1), stop at 40 m
stop at 60 m (1.5 m s -1), [UFP] at 60 m = 100% at 40 m
stop at 60 m (1.5 m s -1), [UFP] at 60 m = 90% at 40 m
stop at 60 m (1.5 m s -1), [UFP] at 60 m = 80% at 40 m
Exposure level of transit-users to UFP around intersections
Set two UFP zones: within ± 20 m of the intersection (high UFP) vs. around (40 and 60 m) (low UFP).
Transit-user’s behavior includesdisembarking, walking, crossing the intersection, waiting for a bus; assuming three pedestrian walk speeds: 0.5 (slow), 1.0 (comfortable), and 1.5 m/s (normal)
Simple time-duration model to simulate exposure reductions when the bus-stop is moved from 20 m to 40 m (or 60 m) from the intersection:
78
Summary• Mobile data offer many advantages as a data collection
strategy, but taking advantage of their high spatial resolution presents special challenges to data processing.
_____________________________________________• A microdynamics model can successfully reproduce
pollutant concentrations at a highly detailed level, and can be used to probe the impacts of particular events and traffic control strategies.
_______________________________________________• Exposures of transit users could be lowered substantially
by moving bus stops from ~ 20 m from the intersections to ~ 40 m.
80
Summary• In the special case of street canyons, the simple building
weighted height/street width aspect ratio can be used to estimate the impact of different designs on pollutant concentrations. This has been verified in many studies.
_______________________________________________
81
Summary• For more complicated urban configurations typical of
California cities, morning concentrations (which are typically highest) appear to be governed by the aerial aspect ratio, the area-weighted building height divided by the fraction of open space, over a several block area.
_____________________________________________• Afternoon concentrations appear to be governed by the
vertical component of the turbulence, which is enhanced by building heterogeneity.
______________________________________________• This part of the study is novel in both its design and
analyses, and is worthy of further investigation.
82
Summary for Planners: Built environment and traffic management design characteristics
that influence near-roadway exposures to vehicular pollutionManagement Suggested
DirectionApprox. Size of
EffectAtmospheric
Conditions & NotesAreal aspect ratio (Aarea)Aarea combines building area-weighted height, building footprint, and the amount of open space.
Lower building volumes and more open space result in lower pollutant concentrations.
Up to approximately a factor of three.
Important under calm conditions (in the mornings at our sites). Not critical when the atmosphere is unstable.
Building Heterogeneity
Isolated tall buildings result in lower concentrations than homogeneous shorter or higher buildings with similar volume.
Up to approximately a factor of two.
Important under unstable conditions with moderate winds (afternoons at our sites). Not critical when the atmosphere is stable.
83
Summary for Planners: Built environment and traffic management design characteristics
that influence near-roadway exposures to vehicular pollutionManagement Suggested
DirectionApprox. Size of
EffectAtmospheric
Conditions & NotesSpecifically within Street Canyons
Ratio of area weighted building height to street width, should be limited to reduce exposure to elevated concentrations of traffic emissions.
Traffic flow Lower traffic flow is better, controlling for fleet mix.
At a given location, concentrations are roughly proportional to traffic flow.
84
Summary for PlannersManagement Suggested Direction Approx. Size of
EffectAtmospheric
Conditions & NotesTraffic Management
Fewer stops and smaller queues reduce emissions and elevated concentrations around intersections
Cannot estimate from our data
Concentrations depend on emissions, micro-scale turbulence, dispersion, transport from nearby streets, and other factors
Bus Stop Siting Further from the intersection is better, but improvements diminish within several tens of meters, depending on built environment (block length, queue length, etc.
Up to approximately a factor of two, but generally 30 –50%.
Measurements are available for calm to moderate winds, when the effect is likely to be strongest.
Sensitive uses near highways
Further is better, but under normal daytime conditions 500 feet is sufficient. If there are consistent nocturnal surface inversions, much longer distances are recommended.
Up to a factor of four or more.
Much more important during surface inversions, which usually occur during night and can persist through mid-morning.
85
The People Who Really Did the Work:Dr. Wonsik ChoiDilhara RanasingheProf. J.R. DeShazoLisa Wong
CARB:Dr. Kathleen KozawaSteve MaraDr. Toshi Kuwayama
Supported by the California Air Resources Board, and the National Science Foundation
86
Dr. Shishan HuKaren BunavageDr. Rodrigo Siguel (UTAM, Chile)Prof. Arthur WinerProf. Mario GerlaProf. Brian Taylor
Si Tan Prof. Akula Venkatram