Science of the Total Environment 466–467 (2014) 277–286
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Evaluating vehicle re-entrained road dust and its potential to deposit to
Lake Tahoe: A bottom-up inventory approach
Dongzi Zhu ⁎, Hampden D. Kuhns, John A. Gillies, Alan W. Gertler
Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, United States
H I G H L I G H T S
•
•
•
•
•
Bottom-up approach provides more geographic accurate traffic pollutant emission.
Geocoded spatial and temporal road dust PM10 emission factor in a road network
Integrated with TDM traffic activities into a regional emission inventory database
GIS tools to analyze each road link position to the lake
Combined dispersion modeling to estimate annual PM10 deposition to Lake Tahoe
a r t i c l e
i n f o
Article history:
Received 5 April 2013
Received in revised form 2 July 2013
Accepted 2 July 2013
Available online xxxx
Editor: P. Kassomenos
Keywords:
Road dust
PM10
Particle deposition
Water clarity
GIS
Lake Tahoe
a b s t r a c t
Identifying hotspot areas impacted by emissions of dust from roadways is an essential step for mitigation.
This paper develops a detailed road dust PM10 emission inventory using a bottom-up approach and evaluates
the potential for the dust to deposit to Lake Tahoe where it can affect water clarity. Previous studies of estimates of quantities of atmospheric deposition of fine sediment particles (“FSP”, b 16 μm in diameter) to the
lake were questioned due to low confidence in the results and insufficient data. A bottom-up approach
that integrates measured road dust emission factors, five years of meteorological data, a traffic demand
model and GIS analysis was used to estimate the near field deposition of airborne particulate matter
b 16 μm, and assess the relationship between trip location and the potential magnitude of this source of
atmospheric deposition to the lake. Approximately ~ 20 Mg year−1 of PM10 and ~ 36 Mg year−1 Total
Suspended Particulate (TSP) from roadway emissions of dust are estimated to reach the lake. We estimate
that the atmospheric dry deposition of particles to the lake attributable to vehicle travel on paved roads is
approximately 0.6% of the Total Maximum Daily Loadings (TMDL) of FSP that the lake can receive and still
meet water quality standards.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
Road traffic emissions are the major source of many pollutants like
carbon monoxide (CO), oxides of nitrogen (NOx), and particulate
matter (PM) in many areas of the US. Air quality modelers rely on
data from national inventories like the US EPA's National Emission
Inventory (NEI), to develop top-down approaches to estimate highway vehicle emissions. These emission data are often used as inputs
to regional scale models like the Community Multi-scale Air Quality
(CMAQ) model. These models lack a detailed spatial distribution of
pollutant concentrations and emissions are allocated to grid cells as
spatial surrogates (Cook et al., 2008). Alternatively, researchers
(Cohen et al., 2005; Kinnee et al., 2004) have used a bottom-up
⁎ Corresponding author. Tel.: +1 775 674 7086; fax: +1 775 674 7007.
E-mail address: zhu@dri.edu (D. Zhu).
0048-9697/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.scitotenv.2013.07.017
approach to provide emission data for inputs to local scale dispersion
models like AERMOD, ADMS or CALINE. The emission data combined
with geocoded traffic activity data from regional travel demand
models (TDM) can be used to develop more accurate vehicle emissions as a function of location. Integrating real-world measured or
modeled vehicle emission factors with TDMs allows more accurate
local emission inventories to be developed. This bottom-up inventory
approach combined with dispersion models can provide information
on spatial variation of pollutant concentrations at receptor sites, e.g.,
to identify hot-spot areas that are most impacted by the traffic activities in an urban area. Wu et al. (2005) developed a GIS-based methodology for producing improved roadway emissions and used the
emissions as inputs for the CALINE4 dispersion model simulation.
Lobdell et al. (2011) applied a bottom-up methodology as presented
in Cook et al. (2008) to predict mobile source emissions (NOx and
PM2.5) in New Haven, CT to assess the public health impact. Beevers
et al. (2012) combined local-scale (ADMS) and regional (CMAQ)
278
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
models to predict hourly pollutant concentrations for London at a
spatial scale of 20 by 20 m. In that study, the authors used hourly traffic emissions (NOx and O3) based on a time series of traffic counts
from 37 automatic sites in London. Jensen et al. (2009) estimated
vehicle emissions in street canyons in New York City based on a
bottom-up approach using traffic data from a travel demand model
and the US EPA MOBILE6.2 emission factors for NOx and PM2.5 for
an intra-urban exposure assessment. Jensen et al. (2009) also observed that resuspended road dust is dominated by coarse particles
(PM2.5–10), however, the resuspension PM10 emission factor was not
included in MOBILE6.2 hence the impact of traffic-generated PM10
(i.e., road dust) was not assessed in that study. Seigneur et al.
(2000) reviewed ten Eulerian air quality models for PM and concluded that all existing models present deficiencies in their simulation of
PM concentrations. Therefore, there is a need to model near-source
deposition of resuspended PM10 from roadways, for example, in
areas where it may be causing an air quality issue or contributing to
the degradation of an important component of the local environment.
This paper describes a case study that applies the bottom-up approach to estimate regional road dust emissions, and the magnitude
of the potential deposition of this source of PM to Lake Tahoe,
which can in turn affect the water clarity of this important alpine lake.
Among the factors contributing to the decline in water quality at
Lake Tahoe are inputs of nitrogen, phosphorous, and sediment into
the lake. Cliff and Cahill (2000) reported that atmospheric deposition
accounts for approximately 55% of the nitrogen (N) and 15% of the
phosphorous (P) load into the lake. In the Lake Tahoe Atmospheric
Deposition Study (LTADS), Dolislager et al. (2012) estimated that approximately 160 Mg of PM2.5, 230 Mg of PM10, and 590 Mg of TSP
(Total Suspended Particulate) were deposited into the lake per year
via this pathway. Estimates of the atmospheric deposition flux of
fine particles into the lake were rated by Roberts and Reuter (2007)
as having the lowest confidence in accuracy due to high uncertainty
in the measurements and an insufficient amount of data. With the
LTADS deposition data incorporated into the Lake Tahoe Total Maximum Daily Load (TMDL) Program (CWB and NDEP, 2010), urban upland loading (i.e., storm water runoff) was attributed to be the main
source (72%) of fine sediment particles (“FSP”, b 16 μm in diameter)
to the lake with atmospheric deposition accounting for 15% of FSP
load in the lake. TMDL describes the maximum load of a pollutant
the lake can receive and still meet water quality standards.
Recent studies have demonstrated that pollutants from the transportation system in the Lake Tahoe basin adversely affect air quality
(Gertler et al., 2006). Airborne particles from re-entrained road dust
have the potential to deposit to the lake and hence decrease water
clarity. Due to the long period when snow can fall in the area
(November–April), traction control sand is periodically applied on the
road surfaces in the basin. When the road surfaces dry, resuspended
dust caused by vehicle movement is a source of PM10 in this region
(Engelbrecht et al., 2009). As a supplementary analysis project of
LTADS, Engelbrecht et al. (2009) analyzed the filters collected as part
of LTADS using Principal Component Analysis. They reported that
resuspended road dust accounts for 38% to 55% of the PM10 in the
Lake Tahoe basin. Similar attribution level (86%) of PM10 to a road
dust source has been reported by Beltran et al. (2012) in Bogota, Colombia. In that study, the spatial disaggregation of PM10 showed that
non-exhaust PM10 emissions are higher in the south-west of the city,
an area with poorly maintained roads, higher volumes of vehicles, and
with the highest levels of particle concentrations in the city. Amato et
al. (2009) studied the spatial pattern of PM10 in road dust deposited
in Barcelona, Spain and found that minerals like SiO2, Al2SO3, and Ca
account for ~40% of PM10. Zhu et al. (2009) measured seasonal and spatial variations of road dust emissions around Lake Tahoe using the
TRAKER vehicle (Etyemezian et al., 2003), a mobile platform that
measures the dust suspended behind the vehicle's front tires. Data
from Zhu et al.'s (2009) TRAKER study is used in this study to construct
a representative road dust emission factor (g-PM10 per vehicle kilometer traveled [g-PM10 VKT−1]) database for vehicles traveling on different road types and road maintenance jurisdiction areas in the Tahoe
basin. The road dust emission factor database of Zhu et al. (2009) was
integrated with a regional travel demand model, which has information
on the spatial distribution of vehicle activities. The model gives geocoded information of annual average daily traffic (AADT) and vehicle
kilometers traveled (VKT) for each road link in the Tahoe basin. The
modeling approach developed here can provide a more accurate estimate of the airborne PM10 road dust deposition component for the
Lake Tahoe TMDL model, leading to an improved estimate of total fine
particle loading to the lake.
2. Methodology
2.1. Road dust emission factor database
In the one-year TRAKER survey around Lake Tahoe (Zhu et al., 2009),
the 41-section, 120 km (75 miles) route was a combination of different
road types (primary, secondary, and tertiary) and covered all road
maintenance organizations in the basin, including Washoe County,
Carson City, and Douglas County in Nevada, and Placer and El Dorado
counties in California. TRAKER measurements of PM10 were calibrated
against PM10 flux measured downwind of a paved road using flux
towers (Zhu et al., 2009). With this calibration TRAKER signals are
converted into emission factors (EFs) in g-PM10 VKT−1. Gillies et al.
(2005) reported that the vehicle road dust emission factor is linearly related to vehicle speed and weight. The TRAKER vehicle weighs 3.1 Mg,
corresponding approximately to the average weight (3.0 Mg including
all vehicle types) of the US vehicle fleet based on the default travel
distance-weighted vehicle mix used in the MOBILE6 emission model
(US EPA, 2002). The weight of TRAKER is also close to the fleet average
in the Tahoe region as Kuhns et al. (2004) reported that cars and SUVs
dominate (~80%) the on road vehicle fleet in the basin, with heavy
duty trucks accounting for ~2% of the fleet. During the survey TRAKER
traveled at the posted road speed limits. The Tahoe TRAKER study was
used to construct a representative road dust emission factor database
for vehicles traveling on different road types and jurisdiction areas in
the basin. The road dust PM10 emission factors in 41 road sections
were grouped into two seasons (December to April is designated as
winter and the rest of the year is designated as summer) for five
counties and three road types (Table 1).
Table 1
Road dust PM10 emission factor (EF) in terms of grams per vehicle kilometer traveled
(g VKT−1) based on county and road type from one-year long TRAKER measurement
campaign.
County
Washoe
Washoe
Washoe
Carson
Douglas
Douglas
Douglas
El Dorado
El Dorado
El Dorado
Placer
Placer
Placer
Road type
Primary
Secondary
Tertiary
Primary
Primary
Secondary
Tertiary
Primary
Secondary
Tertiary
Primary
Secondary
Tertiary
Winter
daily
average EF
Winter EF
standard
deviation
Summer
daily
average EF
Summer EF
standard
deviation
(g VKT−1)
(g VKT−1)
(g VKT−1)
(g VKT−1)
0.30
0.62
1.55
0.24
0.27
1.00
1.89
0.74
2.02
1.38
0.61
1.74
3.70
0.09
0.16
0.12
0.11
0.07
0.68
1.99
0.25
1.83
0.16
0.15
1.53
3.70
0.05
0.15
0.59
0.04
0.04
0.27
0.50
0.16
0.55
1.17
0.15
0.65
1.65
0.01
0.06
0.34
0.02
0.03
0.16
0.31
0.14
0.57
0.48
0.04
0.28
1.65
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
279
the categories: Shrubs, Open Trees, and Dense Trees, with the median
tree coverage in each category being 10%, 35%, and 75%, respectively.
For every shortest path oriented towards the lake for the 7235 traffic
segments, the fraction of each type of vegetation coverage was estimated using ArcGIS analysis tools. For the 7235 traffic points, the average distance to the lake is 2100 m, with a median distance to the
lake of 890 m and a range from 4 m to 18,069 m. Fig. 4 shows a sample GIS overlay with shortest paths defined and the vegetation coverage en route to the Tahoe shoreline.
2.2. Travel demand model
The Tahoe TransCAD travel demand model is a micro-simulated,
tour-based model that reflects resident, seasonal resident, and visitor
travel behavior in the Lake Tahoe basin (Parsons Brinckerhoff, 2007).
The TransCAD road network is built with a Geographic Information
System (GIS). The TransCAD model output provided the AADT, VKT
and the link length of each of the 7235 traffic sections in the basin.
To study the diurnal variation of traffic flow in the basin, road traffic
counters were deployed near Bourne Meadow and Rabe Meadow on
US-50 (primary road; April, 2009), Village Blvd. (secondary road;
June, 2010) in Incline Village, and in Tahoe City on SR-28 (primary
road; May, 2010). Traffic volumes peak in the morning and midday
and are lowest at night on all the roads (Fig. 1). We assumed that
there was no significant seasonal variation in this diurnal traffic
flow pattern for this study. To reflect this flow pattern, TransCAD
model AADT data were categorized into four distinct daily time
frames: AM peak, midday, PM peak, and late night. Hourly traffic
volume in each frame was adjusted for the variation in traffic flows.
2.4. Meteorological network data
Five years (2005–2009) of hourly wind speed and wind direction
data around Lake Tahoe were obtained from the University of California,
Davis's REMOTE network (http://remote.ucdavis.edu/). The REMOTE
project had six meteorological stations around the lake at: Cave Rock,
Timber Cove, Rubicon, Sunnyside, USCG, and Tahoe Vista (Fig. 5). The
yearly wind roses for these stations are shown in Fig. 6.
2.5. PM deposition model
2.3. PM transport route and vegetation density map
Based on the work of Noll and Aluko (2006), we have calculated parameters for a first-order near-source deposition model that accounts
for large particle deposition within the first kilometer of emissions. At
a constant height above the ground, downwind concentrations are assumed to follow an exponential decay:
The roadway network in the Tahoe basin is divided into 7235 individual segments, each modeled as an equivalent finite line source of
PM10 centered at the segment's midpoint (Fig. 2). The shortest pathway (L) from the segment center to the lake shore was acquired using
the GIS functions: the Add Line script, the XTools script, and the Grips
and Shapes script. Landscapes (i.e., density of trees and buildings) can
play a large role in attenuating concentrations of air pollutants before
they reach the lake. The Tahoe Basin Existing Vegetation Map V4.1
(Dobrowski et al., 2005) was imported into ArcGIS (ESRI, Redlands,
CA) to run a vegetation density analysis. A map was created through
the use of digital remote sensing analysis that produced a raster
image containing a cover type for each pixel in the image, which corresponds to a diameter of 1 m. This provided the means to estimate
the percentage of surface covered by trees (Fig. 3). Dominant tree
species in the Lake Tahoe basin are coniferous trees including Jeffrey
pine (Pinus jeffreyi), Lodgepole pine (Pinus contorta) and Western
white pine (Pinus monticola), Ponderosa pine (Pinus ponderosa),
white fir (Abies concolor), and red fir (Abies magnifica); however, deciduous trees such as Aspen may form stands close to the lake. Based
on the different tree coverage classes, the vegetation was divided into
V X
d
− UH
C ðxÞ ¼ C 0 e
ð1Þ
where
C(x)
C0
Vd
X
U
H
is the particle concentration at x meters of horizontal distance from source
is the particle concentration at source
is the deposition velocity, cm s−1
is the meters of horizontal distance from the source
is the horizontal wind velocity, cm s−1
is the injection height of resuspended particle source and
was assumed to be 200 cm.
Cowherd et al. (2006) studied the removal rates of fugitive dust
by different vegetation types impacted by dust plumes created by
Rabe Meadow Hwy50 04-2009
Tahoe City SR28 05-2010
Incline Village Secondary road 06-2010
600
80
Primary road vehicle volume
60
400
50
300
40
30
200
20
100
Secondary road vehicle volume
70
500
10
0
0
Time
Fig. 1. Traffic volume variation measured at two primary roads and one secondary road in the Lake Tahoe basin.
280
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
Fig. 2. TransCAD modeling of midpoints of road links in Tahoe City, CA.
military vehicles traveling on unpaved roads. They found that PM
emission rate (E) similarly exhibits an exponential decay:
−kx
EðxÞ ¼ Eð0Þe
ð2Þ
where the exponent k is a constant term that varies with different
vegetation barriers and x is downwind distance from source. Estimated k values of 0.035 m−1 for dense trees, 0.0175 m−1 for long grass,
and 0.0035 m−1 for short grass are based on their field results.
Eq. (2) is similar to Eq. (1); the k in Eq. (2) should be linearly related
with Vd in Eq. (1). This implies that Vd should increase with increasing
k as vegetation density increases.
Zhu et al. (2012) made PM deposition measurements for five different surfaces (dense long grass, grass cut, sagebrush, desert grass,
no vegetation) and found that particle deposition velocity (Vd)
increases with particle size and surface roughness. The Vd values for
the three types of vegetation for PM2.5, PM10, and PMlarge (particles
N 10 μm) are based on near-source PM deposition measurements at
Ft. Riley, KS, and Hanford, WA (Zhu et al., 2012). Vd values were selected as 10.8 cm s−1, 6.3 cm s−1and 3.6 cm s−1 for PMlarge, PM10,
and PM2.5 for Dense Trees respectively; 8.5 cm s−1, 5.1 cm s−1 and
2.1 cm s−1 for PMlarge, PM10, and PM2.5 for Open Trees respectively;
4.6 cm s−1, 3.2 cm s−1 and 1.7 cm s−1 for PMlarge, PM10, and PM2.5
for Shrubs, respectively. Eq. (1) was applied to the Tahoe area under
certain wind speeds (U) and wind directions. The angle between
the shortest path (which is perpendicular to the road) and wind direction is defined as the attacking angle θ (both the shortest path
and wind are expressed as true-north based azimuth). Only an
angle θ between − 90° and 90° is treated as a favorable wind direction
for a specific traffic point. The actual dust travel distance X along the
wind direction can be expressed as X = L / cosθ. L is the shortest distance to the lake for each traffic point and is a sum of L1, L2, and L3,
which are the corresponding distances covering the Shrubs, Open
Trees, and Dense Trees.
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
281
Fig. 3. Tree coverage (percentage) in the Lake Tahoe basin.
All 7235 traffic sections were assigned to their nearest meteorological station. The mass of PM reaching the lake as a result of vehicle
travel on the 7235 road segments in the basin after attenuation by the
vegetation is estimated as:
PM ¼
Xn
Traffic volume Link length
−1
ðV d1 L1 þ V d2 L2 þ V d3 L3 Þ
exp
UH cos θ
i¼1 EF
i
ð3Þ
where
N
U
H
Vd1
Vd2
Vd3
is the number of traffic segments
is the horizontal wind velocity, cm s−1
is the injection height of resuspended particle source
(200 cm)
is the PM deposition velocity under Shrubs, cm s−1
is the PM deposition velocity under Open Trees, cm s−1
is the PM deposition velocity under Dense Trees, cm s−1.
As mentioned above, EF has a winter and a summer factor, traffic
volume is variable throughout the day, and the wind speed U and
angle θ measured from the perpendicular to the road all change hourly
based on the nearby meteorological station data. Thus, the five-year average (2005–2009) of PMfine, PM10, and PMlarge reaching the lake can be
calculated by integrating Eq. (3) through time. The vegetation roughness seasonal change (loss of leaves in the fall for deciduous trees)
was not considered in the PM deposition model. The Jeffrey pine–
white fir group is the most widespread forest type in the Lake Tahoe
basin, with deciduous trees like Aspen accounting for ~3% of forest composition (Beaty and Taylor, 2008).
3. Results and discussion
Using Eq. (3) annual average PM10 deposition to the lake is estimated
at 20 ± 10 Mg year−1, PMlarge deposition to the lake is estimated at
15 ± 7 Mg year−1, and PMfine (PM2.5) deposition to the lake ranges
282
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
Fig. 4. Sample vegetation coverage of three vegetation type (Shrubs, Open Trees and Dense Trees displayed as white, gray, and black, respectively) and shortest path (dotted lines in
the figure) from traffic points in the Tahoma region to the Lake. The black dots are 1/10 of the length and used to calculate the length coverage fraction in each vegetation category.
from 0.23 ± 0.12 to 3.0 ± 1.5 Mg year−1. The PMlarge and PM2.5 EF
values in Eq. (3) are based on roadside PMlarge concentration measurements, which is 137% of PM10 and the PM2.5 concentration is 1.4% to
18% of the PM10 concentration as measured in the Tahoe area by
Kuhns et al. (2007) and Gertler et al. (2006), respectively. The annual
total PM deposition to the lake from the vehicle travel in the Tahoe
Fig. 5. Meteorological station locations around Lake Tahoe for the UC Davis REMOTE
network.
basin is estimated at 36 ± 12 Mg year−1. The annual PM10 deposition
(to the lake) is approximately 2% of the ~1040 Mg of PM10 resuspended
by vehicles traveling on ~1300 km of roadways in the basin, with an annual daily VKT estimated to be ~2.3 million. The annual total PM deposition (to the lake) is approximately 1.4% of ~2465 Mg total PM
resuspended by vehicles traveling on paved roads in the basin. Table 2
summarizes the total PM (fine, coarse, and large) deposition to the
lake with contributions from different counties in the basin.
These results (Table 2) indicate that 99% of resuspended road dust is
captured by the vegetation or deposited on the pathway leading to the
lakeshore; however, some of that dust could still be transported into the
lake via water runoff following deposition to the land surface. While this
study concentrated on the atmospheric deposition of fugitive road dust,
water runoff carrying the particles of road dust origin can contribute
significantly to the FSP in the lake. CWB and NDEP (2010) report that
in the Lake Tahoe basin the total suspended sediment (TSS) event
mean concentrations (EMCs) from impervious roadways contribute
~30% of TSS loading in storm water transport to the lake. The contribution for upland fine sediment loadings from roadways is also reported
to be ~30% (CWB and NDEP, 2010). Considering that upland water runoff contributes ~72% of FSP loading to the lake (CWB and NDEP, 2010),
water runoff from roadways accounts for ~22% of FSP loadings to the
lake. This suggests street sweeping or other collection efforts should
be made to reduce the total PM load to the lake before it is available
for transport either by resuspension or storm water runoff. Our model
results indicate that 90% of total PM deposition occurs within 500 m
of the lake. Thus, mitigation efforts (e.g., street sweeping) should be
taken to ensure that these roads are kept as clean as possible, thus reducing the available TSS.
For the winter period from December to April in the basin, the PM10
resuspended and subsequently deposited to the lake averages
14 Mg year−1. Approximately 66% of this PM10 can be attributed to
the traction control material applied on the roads (and some track-out
soil brought onto secondary roads by vehicles). Our analysis indicates
that El Dorado County, CA (which includes the City of South Lake
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
USCG (Tahoe City)
Tahoe Vista
Sunnyside
Cave Rock
Rubicon
Timber Cove (SLT)
Fig. 6. Wind rose map from 1-year (2006) monitoring data for the six meteorological stations around Lake Tahoe.
283
284
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
Table 2
Total PM (fine, coarse, large) deposition contributions and vehicle kilometers traveled
(VKT) from different counties around the lake.
County
El Dorado, CA
Douglas County, NV
Placer County, CA
Washoe County, NV
Carson City, NV
Total
Total PM deposition to lake
(Mg year−1)
Annual
Winter
Annual ratio
21
7.2
5.7
0.91
0.005
36
12
5.9
4.0
0.62
0.004
22
61%
20%
16%
2.6%
0.0%
VKT
VKT ratio
1,264,703
345,531
455,463
141,913
11,137
2,218,750
57%
16%
21%
6.4%
0.5%
Tahoe) is the largest contributor (57%) to the total PM deposition to the
lake attributable to resuspended material from roadways as this area
accounts for 51% of the VKT in the basin and is in close proximity to
the lake. Douglas County in NV contributed 21% of the total PM deposition and Placer County, CA, contributed 18% of total PM deposition, with
13% and 25% of basin wide VKT, respectively. Another factor that makes
El Dorado County (including South Lake Tahoe) a dominant contributor
of total PM deposition to the lake is the prevailing south/southwest
wind in that region, which transports the PM towards the lake. Fig. 7
shows the annual PM10 deposition potential in the basin for 7235 traffic
segments, taking into consideration the VKT, seasonal emission factors,
wind speed and direction, distance to the lake, and vegetation barrier
Fig. 7. Gridded annual total PM deposition potential for 7235 traffic segments, taking into consideration the vehicle kilometers traveled (VKT), seasonal emission factors (EFs), wind
speed and direction, distance to the lake, and vegetation barrier density. Each colored group represents 20% of the TSP (Total Suspended Particulate) deposited to the lake, with
classification based on cumulative 20%, 40%, 60%, 80%, and 100% of deposition potential.
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
density. This figure also shows the traffic links with the highest PM deposition potential are located at the near-shore zones, which indicates
that the distance from the road to the lake is a significant factor
impacting the amount of PM deposition to the lake.
As discussed above, the impact of vehicle trips on lake water clarity depends on seasonality, vehicle type, speed, proximity to the lake,
and wind direction and vegetation barriers. Winter and summer seasons have different road dust PM emission factor and traffic volumes.
The estimated year-round population of the Lake Tahoe Basin is approximately 56,000 residents with peaks exceeding 100,000 during
summer months. Generally the highest annual daily traffic volumes
in the basin occur in August.
Analysis by Zhu et al. (2009) shows the wintertime road dust
PM10 emission factor is five times that of the mean summer emission
factor (EF); despite the higher summer seasonal traffic volume. Lower
traffic counts in winter help to reduce the overall impact of
resuspended road dust to the lake. Furthermore, midday (7:00–
15:00) accounts for ~ 70% of daily traffic flow on primary roads and
secondary roads. As Kuhns et al. (2010) and Dolislager et al. (2006)
reported and meteorological data from UC Davis REMOTE sites indicate, in the absence of strong synoptic systems (like winter storms),
the near shore wind has a consistent diurnal pattern. During daytime,
beginning with sunrise, on-shore winds dominate pushing peak
emissions away from the lake, then after sunset and until the following morning off-shore winds dominate. The reduced nighttime traffic
volume fortuitously reduces the direct airborne PM deposition to the
lake even with off-shore winds dominating during that time. Even
though direct airborne transport of PM to the lake is limited by the
aforementioned processes, the high-volume vehicle-resuspended
road dust from the daytime deposits on to road surfaces, curbs, shoulders, and nearby vegetation and soils; these dust deposits may still
enter Lake Tahoe via water runoff or fugitive wind erosion processes,
especially if the dust is deposited back onto road surfaces where it
will be re-entrained again. This track-out effect helps replenish the
silt loadings of the re-suspendable road dust (Kuhns et al., 2010).
Fig. 8 displays cumulative VKT distribution as a function of distance
from the lake in the Tahoe basin. Over 80% of the VKT is within
3000 m of the lake and 70% of the VKT is within 2000 m of the lake.
Thus to mitigate PM10 deposition, street sweeping should be concentrated on roads closest to the lake.
The Tahoe TMDL (CWB and NDEP, 2010) estimates that particles
deposited from the atmosphere account for 15% of the loading of
75 × 1018 particles (1136 Mg based on 66 × 1015 particles per Mg)
(CWB and NDEP, 2008) into the lake. In this study, we estimate that
the mass of sediment reaching the air column over the lake from
paved road dust sources is only 36 Mg due to near-field deposition
Cumulative Distribution of VKT in the Lake Tahoe Basin
Cumulative VKT (%)
1
0.8
0.6
0.4
0.2
285
and the predominance of onshore winds during peak travel times.
The TMDL estimates are modeled deposition fluxes that represent
all sources rather than just paved road dust, which is a subset of all
sources. Although not likely to be a factor in the winter, the magnitude of unpaved road emissions is on the same order as paved road
dust emissions as presented by CWB and NDEP (2008). Another
source of differentiation may be due to the method of the LTADS deposition flux calculation. That model assumes a uniform aerosol concentration based on shoreline measurements across each of the four
lake quadrants. Our results indicate that PMlarge and PMcoarse are
rapidly depleted near their sources. In effect, this greatly reduces
the deposition area of the lake and may account for a large part of
the decrease.
4. Conclusions
Using a bottom-up approach, a GIS analysis integrating road dust
emission factors with geo-coded travel demand model traffic activities was combined with a near-source PM deposition model to estimate the annual dust deposition from paved roadway sources to
Lake Tahoe. This approach provides more geographically accurate information of vehicle activity levels and emissions and improves the
resolution of the spatial gradient of pollutant concentrations in the
dispersion model for an airshed. Five years of wind measurements
from six stations around the basin were linked to each road segment
to account for local transport of resuspended road dust. Diurnal wind
patterns had the fortuitous tendency of blowing road emissions away
from the lake during peak traffic periods. Overall, only ~ 2% of emitted
PM10 and 1.5% of TSP were estimated to reach directly the lake via atmospheric transport and deposition. An integrated assessment of the
deposition potential for road dust emissions to reach the lake was
also conducted. Proximity to the lake, prevailing wind directions,
and traffic patterns played dominant roles in determining which
roads had the greatest potential to contribute fine particles that
could be transported to the lake.
Overall, roads in El Dorado County (and in particular the City of
South Lake Tahoe) had the highest attribution potential (67%) to generate PM from paved roadways that could potentially deposit to the
lake. High VKT in South Lake Tahoe makes this area a major source
of airborne road dust PM10 that can deposit into the lake via atmospheric transport. Incline Village and Tahoe City are estimated to be
minor contributors to lake loading. The Tahoe TMDL model estimates
that atmospheric-deposited particles account for 15% of the lake loading of 75 × 1018 particles (1136 Mg based on 66 × 1015 particles per
Mg). In this study, we estimate that the mass of sediment reaching
the air column over the lake from paved road dust sources to be
only 36 Mg due to near-field deposition and the predominance of onshore winds during peak travel times. Street sweeping or other collection efforts should be made to reduce the total PM load to the
lake before it is available for transport either by vehicle resuspension
process or storm water runoff, since water runoff from roadways contributes 22% of FSP loadings to the lake. Street sweeping or other
strategies to reduce the road dust loading also has the added benefit
of improving ambient PM10 air quality in the Tahoe basin. This
study provides new information on the potential contributions of
fine particulate matter from roadways around Lake Tahoe and can
be used to guide decision making on appropriate emission control
strategies (e.g., street sweeping, anti-icing, reduced VKT) for the
parts of the basin where they will be most effective (i.e., near shore
roadways in Douglas and El Dorado Counties).
0
0
2000
4000
6000
8000 10000 12000 14000 16000 18000 20000
Distance from Lake (m)
Fig. 8. Cumulative VKT distribution as a function of distance to the lake in the Lake
Tahoe basin.
Acknowledgments
This work was funded by the USDA Forest Service via a grant
through the Southern Nevada Public Lands Management Act. We
286
D. Zhu et al. / Science of the Total Environment 466–467 (2014) 277–286
express our appreciation to the University of California, Davis for providing meteorological data and to Mr. Leon Jiang for GIS support.
References
Amato F, Pandolfi M, Viana M, Querol X, Alastuey A, Moreno T. Spatial and chemical
patterns of PM10 in road dust deposited in urban environment. Atmos Environ
2009;9:1650–9.
Beaty RM, Taylor AH. Fire history and the structure and dynamics of a mixed conifer
forest landscape in the northern Sierra Nevada, Lake Tahoe Basin, California, USA.
Forest Ecol Manag 2008;255:707–19.
Beevers SD, Kitwiroon N, Williams ML, Carslaw DC. One way coupling of CMAQ and a
road source dispersion model for fine scale air pollution predictions. Atmos Environ
2012;59:47–58.
Beltran D, Belalcazar LC, Rojas N. Spatial distribution of non-exhaust particulate matter
emissions from road traffic for the City of Bogota — Columbia. Presented at
the 2012 International Emission Inventory Conference, Tampa, FL, August, 2012;
2012. [http://epa.gov/ttnchie1/conference/ei20/session9/dbeltran.pdf].
Parsons Brinckerhoff. Lake Tahoe resident and visitor model — model description and
final results. Prepared for the Tahoe Regional Planning Agency, August 2007; 2007.
Cliff S, Cahill T. Air quality. Lake Tahoe watershed assessment. Albany, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Research Station; 2000.
p. 131–214. [Chapter 3].
Cohen J, Cook R, Bailey C, Carr E. Relationship between motor vehicle emissions of hazardous air pollutants, roadway proximity, and ambient concentrations in Portland,
Oregon. Environ Model Software 2005;20:7–12.
Cook R, Isakov V, Touma J, Benjey W, Thurman J, Kinnee E, et al. Resolving local scale emissions for modeling air quality near roadways. J Air Waste Manage Assoc 2008;58:
451–61.
Cowherd C, Muleski G, Gebhart D. Development of an emission reduction term for
near-source depletion. Proceedings of 15th International Emission Inventory
Conference: “Reinventing Inventories — New Ideas in New Orleans”, New Orleans,
LA, May 15 2006; 2006.
CWB, NDEP. Lake Tahoe TMDL pollutant reduction opportunity report. California Water
Boards (CWB) and Nevada Division of Environmental Protection (NDEP); 2008
[http://www.swrcb.ca.gov/rwqcb6/water_issues/programs/tmdl/lake_tahoe/docs/
presentations/pro_report_v2.pdf].
CWB, NDEP. Final Lake Tahoe total maximum daily load report. California Water Boards
(CWB) and Nevada Division of Environmental Protection (NDEP); 2010 [http://
www.swrcb.ca.gov/rwqcb6/water_issues/programs/tmdl/lake_tahoe/docs/tmdl_rpt_
nov2010.pdf].
Dobrowski S, Greenberg J, Ustin S. Tahoe Basin existing vegetation map v. 4.1. http://
casil.ucdavis.edu/projects/tbevm, 2005.
Dolislager L, Lashgari A, Pederson J, VanCuren T. Lake Tahoe Atmospheric Deposition
Study (LTADS) final report, final report to the Lahontan Regional Water Quality
Control Board, the Nevada Division of Environmental Protection, and the Tahoe
Regional Planning Agency. Prepared by the Atmospheric Processes Research Section.
Sacramento, CA: California Air Resources Board; 2006.
Dolislager LJ, VanCuren R, Pederson JR, Lashgari A, McCauley E. A summary of the Lake
Tahoe Atmospheric Deposition Study (LTADS). Atmos Environ 2012;46:618–30.
Engelbrecht J, Gertler A, VanCuren T. Lake Tahoe Source Attribution Study (LTSAS):
receptor modeling study to determine the sources of observed ambient particulate
matter in the Lake Tahoe Basin: final report. Prepared for USDA Forest Service
Pacific Southwest Research Station by Desert Research Institute. Reno, NV: Desert
Research Institute; 2009.
Etyemezian V, Kuhns HD, Gillies JA, Green M, Pitchford M, Watson J. Vehicle-based road
dust emission measurement: I—methods and calibration. Atmos Environ 2003;37:
4559–71.
Gertler A, Kuhns H, Abu-Allaban M, Damm C, Gillies J, Etyemezian V, et al. A case study
of the impact of winter road sand/salt and street sweeping on road dust
re-entrainment. Atmos Environ 2006;40:5976–85.
Gillies JA, Etyemezian V, Kuhns H, Nikolic D, Gillette DA. Effect of vehicle characteristics
on unpaved road dust emissions. Atmos Environ 2005;39:2341–7.
Jensen SS, Larson T, Deepti KC, Kaufman JD. Modeling traffic air pollution in street canyons in New York City for intra-urban exposure assessment in the US multi-ethnic
study of atherosclerosis and air pollution. Atmos Environ 2009;43:4544–56.
Kinnee E, Touma J, Mason R, Thurman J, Beidler A, Bailey C, et al. Allocation of on-road
mobile emissions to road segments for air toxics modeling in an urban area. Transport Res D Transport Environ 2004;9:139–50.
Kuhns H, Chang M, Chow J, Etyemezian V, Chen L, Nussbaum N, et al. DRI Lake Tahoe
source characterization study: final report. Prepared for California Air Resources
Board. Reno, NV: Desert Research Institute; 2004.
Kuhns H, Zhu D, Gillies J, Gertler A, Etyemezian V, Brown S. Measurement and modeling
of fugitive dust emissions from paved road travel in the Lake Tahoe Basin. Report prepared for US EPA Region 9, San Francisco, CA. Reno, NV: Desert Research Institute;
2007.
Kuhns H, Zhu D, Gillies J, Gertler A, Cliff S, Zhao Y, et al. Examination of dust and air-borne
sediment control demonstration projects. Reno, NV: Desert Research Institute; 2010.
Lobdell DT, Isakov V, Baxter L, Touma JS, Smuts MB, Özkaynak H. Feasibility of assessing
public health impacts of air pollution reduction programs on a local scale: New
Haven case study. Environ Health Perspect 2011;119:487–93.
Noll KE, Aluko O. Changes in large particle size distribution due to dry deposition processes. J Aerosol Sci 2006;37:1797–808.
Roberts D, Reuter J. Lake Tahoe total maximum daily load technical report. California and Nevada, draft report prepared for Lahontan Regional Water Quality Control Board; 2007.
Seigneur C, Pun B, Pai P, Louis J-F, Solomon P, Emery C, et al. Guidance for the performance evaluation of three-dimensional air quality modeling systems for particulate matter and visibility. J Air Waste Manage Assoc 2000;50:588–99.
US EPA. User's guide to MOBILE6.0 mobile source emission factor model. USA: US
Environmental Protection Agency; 2002 [Report No. EPA-420/R-02-001].
Wu J, Funk TH, Lurmann FW, Winer AM. Improving spatial accuracy of roadway networks and geocoded addresses. Trans GIS 2005;9:585–601.
Zhu D, Kuhns H, Brown S, Gillies J, Etyemezian V, Gertler A. Fugitive dust emissions from
paved road travel in the Lake Tahoe Basin. J Air Waste Manage Assoc 2009;59:1219–929.
Zhu D, Gillies JA, Etyemezian V, Kuhns HD, Engelbrecht J, Nikolich G. Influence of surface roughness length on particle deposition. Paper A518, Proceedings of the Air
and Waste Management Association's Annual Conference and Exhibition, San
Antonio, TX, June, 2012; 2012.