In this study, dispersion and deposition of atmospheric mercury (Hg) in Tehran city was simulated using WRF-SMOKE-CMAQ models. The Weather Research and Forecasting (WRF) model was used to simulate the meteorological parameters. For validation of WRF results; the simulated wind speeds and temperatures were compared with the parameters measured at a meteorological station in Tehran city for 11 days (8 days in fall and 3 days in winter) in 2010 - 2011. The correlation coefficient (r) for temperature and wind speed were 0.94 and 0.49, respectively indicating there was good agreement between measured and modeled results. An atmospheric mercury emission inventory was developed using the United Nations Environment Programme (UNEP), the United States Environmental Protection Agency AP-42 (US-EPA AP-42) and related papers. Sparse Matrix Operator Kernel Emissions (SMOKE) was used to allocate the atmospheric mercury emissions to the modeling domain and the Community Multiscale Air Quality (CMAQ) model was used to simulate the concentration and deposition of atmospheric mercury. To validate the results of the CMAQ model, the simulated atmospheric particulate mercury (PHg) concentrations for 11 days were compared with the measured results at two different stations (Bagh Ferdows and Bahman Square) where it was measured by the Tehran Air Quality Control Company (AQCC). Comparison between the results from the modeled and measurements of PHg in fall was better than winter. Concentrations and dry depositions of the various forms of atmospheric mercury were higher in areas closer to mercury stationary emission sources.

1. Simulation of atmospheric mercury dispersion and deposition
in Tehran city
Mohammadamin Vahidi Ghazvini1
& Khosro Ashrafi1
& Majid Shafiepour Motlagh1
& Alireza Pardakhti1
&
Sarmad Ghader2
& Thomas M. Holsen3
Received: 23 November 2019 /Accepted: 9 March 2020
# Springer Nature B.V. 2020
Abstract
In this study, dispersion and deposition of atmospheric mercury (Hg) in Tehran city was simulated using WRF-SMOKE-CMAQ
models. The Weather Research and Forecasting (WRF) model was used to simulate the meteorological parameters. For validation
of WRF results, the simulated wind speeds and temperatures were compared with the parameters measured at a meteorological
station in Tehran city for 11 days (8 days in fall and 3 days in winter) in 2010–2011. The correlation coefficient (r) for temperature
and wind speed was 0.94 and 0.49, respectively, indicating there was good agreement between measured and modeled results. An
atmospheric mercury emission inventory was developed using the United Nations Environment Programme (UNEP), the United
States Environmental Protection Agency AP-42 (US-EPA AP-42), and related papers. Sparse Matrix Operator Kernel Emissions
(SMOKE) was used to allocate the atmospheric mercury emissions to the modeling domain, and the Community Multiscale Air
Quality (CMAQ) model was used to simulate the concentration and deposition of atmospheric mercury. To validate the results of
the CMAQ model, the simulated atmospheric particulate mercury (PHg) concentrations for 11 days were compared with the
measured results at two different stations (Bagh Ferdows and Bahman Square) where it was measured by the Tehran Air Quality
Control Company (AQCC). Comparison between the results from the modeled and measurements of PHg in fall was better than
winter. Concentrations and dry depositions of the various forms of atmospheric mercury were higher in areas closer to mercury
stationary emission sources.
Keywords Atmospheric mercury . Tehran city . WRF-SMOKE-CMAQ models . Numerical simulation . Dispersion and
deposition
Introduction
Mercury is one toxic pollutant among the numerous heavy
metals that can accumulate in biotic tissues. Human exposure
to mercury can be from exposure to contaminated soil, water,
or food, particularly fish. Mercury has a variety of
documented adverse impacts on human health and the envi-
ronment (Wip et al. 2013). Mercury can accumulate in organs,
such as the kidney, liver, and especially in the brain, and may
interfere with digestive, immune, and nervous systems (Rice
et al. 2014).
Mercury is present in the atmosphere in three main forms
including gaseous elemental mercury (GEM), gaseous oxi-
dized mercury (GOM), and particulate mercury (PHg). The
main component of mercury in the atmosphere is GEM (>
80%), which remains in the atmosphere several months to
1 year (Driscoll et al. 2013). GOM and PHg have a much
shorter residence time in the atmosphere due to more efficient
removal by wet and dry deposition (Marrugo-Negrete et al.
2014). Their atmospheric residence time is days to weeks
(Zhou et al. 2019).
A number of studies have been carried out to simulate the
dispersion and deposition of mercury in the atmosphere at
various scales using Lagrangian (RELMAP and HYSPLIT)
or Eulerian approaches (CMAQ). Since 2000, a number of
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s11869-020-00813-x) contains supplementary
material, which is available to authorized users.
* Khosro Ashrafi
khashrafi@ut.ac.ir
1
School of Environment, College of Engineering, University of
Tehran, Tehran, Iran
2
Institute of Geophysics, University of Tehran, Tehran, Iran
3
Department of Civil & Environmental Engineering, Clarkson
University, Potsdam, NY, USA
https://doi.org/10.1007/s11869-020-00813-x
Air Quality, Atmosphere & Health (2020) 13:529–541
/Published online: 1 April 2020

2. Eulerian models of dispersion and deposition of atmospheric
mercury have been developed at regional and global scales.
The differences between these models are often in their for-
mulations and assumptions about atmospheric processes, re-
actions, and behavior (Bullock and Brehme 2002).
The Community Multiscale Air Quality (CMAQ) model is
three-dimensional with an Eulerian approach that is designed
to estimate the concentration of pollutants and their deposition
on a city to continental scale. The pollutants that are simulated
with the standard CMAQ version include tropospheric ozone,
acidic and neutral species, and particles with different compo-
sitions and sizes (EPA 2015).
Air quality models need meteorological parameters gener-
ated by meteorological models as inputs. One of the best me-
teorological models is Weather Research and Forecasting
(WRF) model which is a weather mesoscale numerical pre-
diction system which is designed for atmospheric research and
operational predictions. This model covers a wide range of
meteorological scales ranging from tens of meters to thou-
sands of kilometers. The main equations of all atmospheric
numerical simulation models, including the WRF model, are
based on the principles of mass, momentum, and energy con-
servation (National Center for Atmospheric Research 2017).
In addition to the meteorological parameters, air quality
models need emission inventories to perform simulations.
Emission inventories are allocated by the emission model to
the study domain. The Sparse Matrix Operator Kernel
Emissions (SMOKE) model is one of the emission allocations
models that is used as input for air quality models. The
SMOKE model also has the ability to process a variety of
pollutants including gaseous, particulate, and toxic pollutants
such as mercury (EPA 2013).
Bullock (2000) simulated the annual average ground level
of GEM, GOM, and PHg along with the annual wet and dry
deposition of all forms of mercury using the Regional
Lagrangian Model of Air Pollution (RELMAP) in the USA.
Bullock and Brehme (2002) simulated the total mercury wet
deposition for a month of spring and summer seasons using
CMAQ model in east of USA. Gbor et al. (2007) simulated
the daily average total gaseous mercury (TGM = GEM +
GOM) concentration in the surface layer and daily dry and
wet deposition of different mercury species for 2002 in a do-
main that covered the continental United States and major
parts of Canada and Mexico. They also used the CMAQ mod-
el, but the difference between those results and most other
studies was that they considered mercury emissions from nat-
ural sources in addition to anthropogenic mercury emissions.
Ryaboshapko et al. (2007a) simulated the concentrations of
different forms of atmospheric mercury using seven different
air quality models with various regional, continental, and
hemispheric scales and compared the results with the short-
term measurements (2 weeks) data at 5 different measurement
stations in Europe. Ryaboshapko et al. (2007b) simulated
GEM concentration, Hg wet and dry deposition using eight
different air quality models with various regional, hemispher-
ic, and global scales and compared the results with the long-
term measurements (months to years) data at 11 different mea-
surement stations in Europe. Holloway et al. (2012) simulated
Fig. 1 Location map showing
Tehran city
Air Qual Atmos Health (2020) 13:529–541
530

3. the concentration of different forms of atmospheric mercury in
the Great Lakes Region of North America using WRF/CMAQ
models and compared the results with measurements data at
two urban and rural stations.
This study investigates atmospheric mercury dispersion
and deposition in Tehran city as no previous atmospheric mer-
cury modeling has been performed for this location. The urban
modeling scale used was a fine grid size (1.3 km × 1.3 km),
smaller than that used in other studies (Bullock and Brehme
2002; Gbor et al. 2007; Ryaboshapko et al. 2007a) which were
at larger regional scales (grids of 36 km × 36 km).
Materials and methods
Location and spatial zoning
The greater Tehran province and its surrounding area were
selected as the domain in this study, due to the surrounding
area impacts on air quality of Tehran city. The surrounding
area is mainly pasture lands. Tehran province consists of nine
main cities: Tehran, Rey, Robat Karim, Islamshahr, Shahriar,
Pakdasht, Varamin, Damavand, and Firouzkooh (Fig. 1).
Tehran city has a complex topography due to its location on
the foothills of Alborz Mountains with the elevation ranging
from 900 to 1800 m above sea level. The slope increases from
a gentle slope in the northern parts of the city to a very steep
slope, reaching the summit peaks of the northern mountains.
Geographic Information System (GIS) software was used
to show the location of domain. The cities of Tehran province
with the main and secondary roads, along with 12 stationary
sources of mercury emissions, including cement factories,
power plants, Tehran oil refinery, and Tehran municipal solid
waste (MSW) landfill, are identified on the map (Fig. S1).
The total domain was divided into 493 zones in which 399
zones are located inside the Tehran city, 90 zones are for the
other main cities in the province and surrounding major roads,
and four zones surround the province and emit naturally oc-
curring mercury (Fig. 2).
Emissions
The atmospheric mercury emission factors from the United
Nations Environment Programme (UNEP 2017), United
States Environmental Protection Agency AP-42 (EPA
1998a, b), and the other related publications were used. In
Tehran province, there are six power plants that use natural
gas and gas-oil as fuels. In addition, there is an oil refinery in
Tehran province that uses natural gas, gas-oil, and fuel-oil as
fuels. Since the measured data for atmospheric particulate
mercury was available for only 11 days between November
2010 and February 2011, the atmospheric mercury emissions
were estimated for a whole year from 21 March 2010 to 20
March 2011. The amount of fuel consumed by the power
plants and the oil refinery was based on data (Table S1) from
Iran Statistics of Consumption of Energy (National Iranian Oil
Refining and Distribution Company 21 March 2010–20
March 2011), Iran Energy Balance sheet (Iran Ministry of
Power 21 March 2010–20 March 2011), and Tehran
Province Statistical Yearbook (Management and Planning
Organization of Tehran Province 21 March 2010–20
March 2011).
There are four cement factories in Tehran province, for
which US-EPA AP-42 provided the emission factors of mer-
cury based on the amount of cement produced and the emis-
sion control methods used (EPA 1994). Another stationary
source is the Tehran municipal solid waste landfill. For this
site, atmospheric mercury emissions from landfill gas were
estimated using emission factors from US-EPA AP-42 based
on the amount of landfill gas produced (EPA 1997).
In addition to the stationary point sources, there are minor
mercury emission sources that consume fuels. These sources
Fig. 2 The modeling domain
consists of 493 zones. Each color
represents a city: blue, Tehran;
purple, Damavand; orange,
Firouzkooh; black, Pakdasht;
white, Rey; brown, Varamin; red,
Islamshahr; yellow, Shahriar;
pink, Robat Karim. Gray color
represents the other roads around
the Tehran province and green
color represents the pastures
surrounding the Tehran province
Air Qual Atmos Health (2020) 13:529–541 531

4. include small industries, domestic, and residential power gen-
erating systems, and other small sources. The amount of fuel
used in the cement factories, power plants, and oil refinery
were subtracted from the total fuel consumption in Tehran
province and allocated to the minor mercury emission sources,
which are referred to as “Stationary Area Sources”. The US-
EPA AP-42 emission factors were used to estimate these emis-
sions. The data from the Iran Statistics of Consumption of
Energy, Iran Energy Balance sheet, and Tehran Province
Statistical Yearbook was also used to estimate the fuel con-
sumption from stationary area sources (Table S2).
Mercury is also produced and emitted into the air
from “Other Stationary Area Sources” in Tehran prov-
ince such as brick manufacturing, lamp production and
usage, mercury used in dentistry, and mercury emissions
from thermometers. The atmospheric mercury emission
from mobile sources from all types of vehicles includ-
ing cars, busses, motorcycles, and trucks in Tehran city
was estimated by Shahrokhi (2015) using the Motor
Vehicle Emission Simulator (MOVES) model which is
shown in Table S3. Mercury is also emitted into the air
from natural sources for which emissions were estimat-
ed based on the urban surfaces that usually include
impervious surfaces such as pavements and permeable
surfaces such as soil and green areas. The atmospheric
mercury emission from these sources was calculated and
the annual atmospheric mercury emission from different
sources in Tehran province was estimated. Table 1 states
the annual estimated atmospheric mercury emissions in
the domain for each source.
Modeling with WRF
In this study, WRF model version 3.9 was used. The
FNL (Final) Operational Global Analysis meteorological
data were obtained during the period considered. The
input files to the model have been set up and the sim-
ulation dates prepared based on FNL data entered into
the model. The modeling area consisted of four nested
domains with 36 km × 36 km grid size at the outermost
domain which decreased by one-third proportions and
reached an innermost domain with the grid size of
1.3 km × 1.3 km including Tehran city. The appropriate
physical meteorological parameters for Tehran city
based on the results of the study by Ghader et al.
(2016) were used, which had the best correlation with
the measured meteorological data. Table S4 represents
the physical parameters used in this study.
Allocation with SMOKE
The Inventory Data Analyzer (IDA) format was used to
allocate the atmospheric mercury emissions in the
SMOKE model, version 3.5.1. Time allocations were
considered for modeling mercury release into the atmo-
sphere during the workdays and holidays. For mercury
speciation, the information of Walcek et al. (2003) and
EPA SMOKE User’s Manual (2013) was used (Table 2).
For speciation of mercury from natural sources, the in-
formation of Travnikov (2005) and Ryaboshapko et al.
(2007a) was used which states that atmospheric mercury
emissions from natural sources are as GEM.
Table 2 Emission profiles for
atmospheric mercury from all
sources
Emission source GEM GOM PHg Reference
Fuels combustion 0.5 0.3 0.2 (Walcek et al. 2003)
Cement manufacturing 0.75 0.13 0.12 (EPA 2013)
Municipal landfill 0.8 0.1 0.1 (Walcek et al. 2003)
Mobile sources—diesel 0.56 0.29 0.15 (EPA 2013)
Mobile sources—gasoline 0.91 0.086 0.004 (EPA 2013)
Other stationary area sources 0.8 0.1 0.1 (Walcek et al. 2003)
Natural sources 1.0 0.0 0.0 (Travnikov 2005; Ryaboshapko et al. 2007a)
Average 0.76 0.15 0.09 –
Table 1 Atmospheric mercury emissions by emission source
considered in the domain
Type of emission sources Atmospheric mercury emission (kg/year)
Mobile sources 45.8
Stationary area sources 133.9
Cement factories 372.8
Power plants 58.2
Oil refinery 5.5
Waste disposal 0.8
Brick manufacturing 12.1
Lamps 29.2
Thermometers 22.7
Dental amalgam preparation 24.4
Natural sources 434.1
Total 1140
Air Qual Atmos Health (2020) 13:529–541
532

5. The GIS’s output was used by the Special Allocator
tool to determine the geographic coordinates of the do-
main. The domain grid consists of a network of 100 ×
76 cells with 1.3-km cell length. Estimated atmospheric
mercury emissions were allocated for stationary point
sources in the respective zones. Estimated atmospheric
mercury emission for stationary area sources were allo-
cated in zones for each city based on population ratio.
Modeling with CMAQ
The dispersion, deposition, and chemical transport of
mercury were modeled using the meteorological param-
eters and the atmospheric mercury emission inventory,
by the CMAQ model version 5.1. The initial and
boundary conditions of mercury were defined inside
and at four lateral boundaries of the domain, respective-
ly. The default CMAQ boundary conditions were used
to handle the mercury emissions of external sources
from outside the domain. The mercury concentration
distribution was simulated using the implementation of
the CMAQ Chemical Transport Modeling System
(CCTM). The grid specifications used in the meteoro-
logical model were transmitted to the CMAQ by the
Meteorology Chemistry Interface Processor (MCIP)
sub-program.
Measurements
The modeled results were compared with the actual
measured data to validate the accuracy of modeling re-
sults. The Tehran Air Quality Control Company report
(AQCC 2011) for two stations, Bahman Square and
Bagh Ferdows (Fig. 3), that measured the particulate
mercury using high volume samplers for 24 h was used
to validate the results obtained from the model.
Atmospheric particulate mercury measurements in
AQCC report was performed for 11 days (8 days in
fall, 2010, and 3 days in winter, 2011) in Tehran city
which atmospheric PHg concentrations is presented in
Table S5. It is important to mention that there are no
measured data available for gaseous mercury in Tehran
city due to lack of measurement sampling equipment.
Results
WRF results
The WRF model was used to generate meteorological
parameters during the 11 days of simulation. The first
8 days of the period are in the fall and the last 3 days
in the winter season. In order to ensure the accuracy of
the modeling results, the temperature and wind speed
modeled results were compared with Mehrabad meteo-
rological synoptic station data (Fig. 4).
Mehrabad meteorological synoptic station reports me-
teorological parameters every 3 h; therefore, 8 sets of
data were obtained during the day. For a more accurate
comparison, the model results were also used for exact-
ly the same hours. WRF reports temperature at 2 m
above the ground level (T2) and wind speed/direction
at 10 m above the ground level (U10, V10). The wind
speed value (W10) was determined using formula (1).
Figure 5 shows the daily correlation between measured
and modeled data for temperature and wind speed pa-
rameters over 11 different days during the simulation
period.
W10 ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
U10
2
þ V10
2
p
ð1Þ
Fig. 3 Locations of stations in
Tehran city. A circle represents
the particulate mercury measuring
stations and square represents the
meteorological synoptic station
Air Qual Atmos Health (2020) 13:529–541 533

6. The results for the temperature parameter show a
good agreement between the model and the measure-
ment values with a correlation coefficient equal to
0.94. The correlation between the measured data and
the modeled data for wind speed is not as good, with
a correlation coefficient of 0.49. Since the Mehrabad
meteorological synoptic station assumes a wind speed
below 2 m/s to be equal to 0, the difference between
the modeled and measured wind speed is greater.
SMOKE results
The SMOKE model was run for each day of the simu-
lation period separately. Fig. S2 shows the output of
-5
0
5
10
15
20
2010-11-08
2010-11-08
2010-11-08
2010-11-08
2010-11-13
2010-11-13
2010-11-13
2010-11-13
2010-11-14
2010-11-14
2010-11-14
2010-11-14
2010-11-15
2010-11-15
2010-11-15
2010-11-15
2010-11-28
2010-11-28
2010-11-28
2010-11-28
2010-11-29
2010-11-29
2010-11-29
2010-11-29
2010-12-05
2010-12-05
2010-12-05
2010-12-05
2010-12-06
2010-12-06
2010-12-06
2010-12-06
2011-01-31
2011-01-31
2011-01-31
2011-01-31
2011-02-06
2011-02-06
2011-02-06
2011-02-06
2011-02-14
2011-02-14
2011-02-14
2011-02-14
Temperature
(°C)
Measured Modeled
(a)
0
1
2
3
4
5
6
7
2010-11-08
2010-11-08
2010-11-08
2010-11-08
2010-11-13
2010-11-13
2010-11-13
2010-11-13
2010-11-14
2010-11-14
2010-11-14
2010-11-14
2010-11-15
2010-11-15
2010-11-15
2010-11-15
2010-11-28
2010-11-28
2010-11-28
2010-11-28
2010-11-29
2010-11-29
2010-11-29
2010-11-29
2010-12-05
2010-12-05
2010-12-05
2010-12-05
2010-12-06
2010-12-06
2010-12-06
2010-12-06
2011-01-31
2011-01-31
2011-01-31
2011-01-31
2011-02-06
2011-02-06
2011-02-06
2011-02-06
2011-02-14
2011-02-14
2011-02-14
2011-02-14
Wind
Speed
(m/s)
Measured Modeled
(b)
Fig. 4 Daily comparison between modeled and measured temperature (a) and wind speed (b) for the Mehrabad synoptic station
Air Qual Atmos Health (2020) 13:529–541
534

7. SMOKE model for atmospheric mercury emissions allo-
cated to each mercury species for 08.11.2010 at 13:00
UTC. The SMOKE model provides the emission rates
per hour for each form of mercury during the day with
GEM emissions being the highest and PHg being the
lowest.
y = 0.95x + 0.18
r = 0.94
-5
0
5
10
15
20
-5 0 5 10 15 20
Modeled
(°C)
Measured (°C)
(a)
y = 0.29x + 0.91
r = 0.49
0
1
2
3
4
5
6
7
0 1 2 3 4 5 6 7
)
s
/
m
(
d
e
l
e
d
o
M
Measured (m/s)
(b)
Fig. 5 Daily correlation between modeled and measured temperature (a) and wind speed (b)
Fig. 6 Atmospheric concentrations of PHg (a), GEM (b), and GOM (c) in the total domain for 28.11.2010 at 19:00 UTC. The boundaries of Tehran city
were used
Air Qual Atmos Health (2020) 13:529–541 535

8. CMAQ results
Dispersion
The CMAQ model was run for each day of simulation period
separately. Figure 6 shows the output of CMAQ model for the
atmospheric concentrations of each species of mercury for
28.11.2010 at 19:00 UTC.
As shown in Fig. 6, atmospheric mercury concentrations in
the southeastern part of Tehran city are higher than in the other
areas. This is due to the presence of the Tehran Cement
Factory in that area which has the largest proportion of annual
cement production among the four existing cement factories
in Tehran province indicating that it is the most polluting
source of atmospheric mercury in Tehran city.
The total amount of atmospheric particulate mercury per
hour during each day was obtained for two measuring stations,
Bagh Ferdows and Bahman Square. The daily mean values of
atmospheric PHg were compared with the measured results at
these stations in Fig. 7. Since there is not any atmospheric PHg
Fall Winter
Fall Winter
(a)
(b)
Fig. 7 Daily modeled concentration of PHg in Bagh Ferdows (a), and Bahman Square (b) stations. The multiplication sign represents the mean, the box
plots are the 25th and 75th percentiles, the square is the measured data, and the line is the median
Air Qual Atmos Health (2020) 13:529–541
536

9. measured data for days 08.11.2010 and 15.11.2010 at the
Bagh Ferdows station, the measurement values on Fig. 7a
are missing for these days.
The model results for almost all days except 31.01.2011 at
both measuring stations are within the same range. The mea-
surement results at both stations in the last 3 days of the simu-
lation period, namely, 31.01.2011, 06.02.2011, and 14.02.2011,
are higher than the other days of the simulation period
corresponding with the model results. The statistical compari-
son of hourly atmospheric PHg concentration for both measur-
ing stations during the simulation period is listed in Table 3.
In order to verify the accuracy of the results at the two
stations, root mean square error (RMSE) was calculated using
formula (2) which there, Oi and Mi represent the daily average
of measured and modeled atmospheric PHg concentration in
the simulation period, respectively.
Table 3 Summary of statistical parameters for hourly atmospheric PHg concentration during the simulation period
Date Station Mean σ Max Min
08.11.2010 Bagh Ferdows 0.066 0.031 0.13 0.031
Bahman Square 0.153 0.106 0.403 0.03
13.11.2010 Bagh Ferdows 0.072 0.044 0.168 0.03
Bahman Square 0.151 0.085 0.343 0.056
14.11.2010 Bagh Ferdows 0.06 0.024 0.103 0.031
Bahman Square 0.164 0.094 0.328 0.027
15.11.2010 Bagh Ferdows 0.077 0.028 0.14 0.051
Bahman Square 0.174 0.114 0.395 0.027
28.11.2010 Bagh Ferdows 0.09 0.044 0.163 0.042
Bahman Square 0.197 0.103 0.423 0.035
29.11.2010 Bagh Ferdows 0.056 0.019 0.119 0.036
Bahman Square 0.181 0.093 0.322 0.023
05.12.2010 Bagh Ferdows 0.078 0.035 0.145 0.035
Bahman Square 0.196 0.11 0.342 0.032
06.12.2010 Bagh Ferdows 0.096 0.043 0.183 0.054
Bahman Square 0.215 0.145 0.593 0.031
31.01.2011 Bagh Ferdows 0.02 0.02 0.076 0.004
Bahman Square 0.044 0.046 0.149 0.008
06.02.2011 Bagh Ferdows 0.065 0.015 0.104 0.046
Bahman Square 0.187 0.085 0.381 0.065
14.02.2011 Bagh Ferdows 0.072 0.029 0.129 0.04
Bahman Square 0.2 0.134 0.465 0.03
0
0.5
1
1.5
2
2.5
3
2010-11-08
2010-11-13
2010-11-14
2010-11-15
2010-11-28
2010-11-29
2010-12-05
2010-12-06
2011-01-31
2011-02-06
2011-02-14
m
/
g
n
(
n
o
i
t
a
r
t
n
e
c
n
o
c
M
E
G
e
g
a
r
e
v
A
y
l
i
a
D
3
)
Bagh Ferdows Bahman square
(a)
0
50
100
150
200
250
300
2010-11-08
2010-11-13
2010-11-14
2010-11-15
2010-11-28
2010-11-29
2010-12-05
2010-12-06
2011-01-31
2011-02-06
2011-02-14
m
/
g
p
(
n
o
i
t
a
r
t
n
e
c
n
o
c
M
O
G
e
g
a
r
e
v
A
y
l
i
a
D
3
)
Bagh Ferdows Bahman square
)
b
(
Fig. 8 Modeled daily average atmospheric concentration of GEM (a) and GOM (b) in Bagh Ferdows and Bahman Square stations
Air Qual Atmos Health (2020) 13:529–541 537

10. RMSE ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
N
∑
N
i¼1
Oi−Mi
ð Þ2
s
ð2Þ
The RMSE was calculated as 0.28 ng/m3
and 0.46 ng/m3
for Bagh Ferdows and Bahman Square, respectively, indicat-
ing that the model performs better for the Bagh Ferdows sta-
tion since to the values of RMSE are closer to 0. The daily
average concentrations of GEM and GOM at both measuring
stations are shown in Fig. 8. Concentrations of modeled GEM
and GOM at Bahman Square station are higher due to their
proximity to emission sources. At 31.01.2011, the concentra-
tions of GEM and GOM decreased due to precipitation and
wet deposition. The RMSE was calculated as 0.08 ng/m3
and
0.7 ng/m3
for fall and winter days, respectively, indicating that
the model performs better for fall than winter season.
Deposition
Figure 9 shows the output of CMAQ model for the dry depo-
sition of each species of mercury for 13.11.2010 at 05:00
UTC. The dry deposition of GOM is larger than for PHg
and GEM. Mercury dry deposition is also larger closer to the
mercury stationary emission sources. In the southeastern part
of Tehran city, the highest dry deposition was observed due to
the presence of Tehran cement factory which has the highest
rate of atmospheric mercury emission from stationary sources
in Tehran city.
Dry deposition was modeled for the simulation period for
the three forms of mercury. Figure 10 shows the mercury daily
dry deposition and total dry deposition at both measuring sta-
tions. Most dry deposition in the GOM form. At 31.01.2011,
due to the precipitation, GOM dry deposition decreased but
GEM and PHg deposition increased at both stations. In the
winter season, the total dry deposition appears to be higher than
in the fall due to the lower temperatures. Total dry deposition at
Bahman Square station is higher than Bagh Ferdows station
due to the higher mercury concentrations.
Figure 11 shows the output of CMAQ model for the wet
deposition of each species of mercury for 31.01.2011 at 06:00
UTC. The wet deposition of GOM is larger than PHg and
GEM and is dependent on the precipitation patterns.
During the simulation period, only on 31.01.2011 precipi-
tation occurred and the wet deposition was obtained for this
day. Figure 12 shows the hourly wet deposition for three
Fig. 9 Dry deposition of PHg (a), GEM (b), and GOM (c) in the total domain for 13.11.2010 at 05:00 UTC. The boundaries of Tehran city were used
Air Qual Atmos Health (2020) 13:529–541
538

11. forms of mercury and the total wet deposition at both measur-
ing stations. Most of the wet deposition is in the GOM form
and wet deposition of GEM is negligible. The total wet depo-
sition of Bagh Ferdows station is larger than at Bahman
Square station due to more precipitation. Precipitation stopped
at the end of the day in Bahman Square station; therefore, the
wet deposition was 0 during those hours.
Conclusions
In this study, the behavior of mercury in a domain covering
Tehran city was simulated using WRF-SMOKE-CMAQ
models to simulate the atmospheric concentration and deposi-
tion of mercury in Tehran city.
Comparison of modeling and measuring results for at-
mospheric PHg concentration in winter season (the last
3 days of the simulation period) was not very good, which
seems to be due to a problem with the measured data
since it is highly variable compared with the measured
data in the other days of this period (fall season). Model
results in the fall season are closer to the measured data
compared with winter. However, due to the low concen-
tration of atmospheric PHg, these differences between
measured and model values seem to be inevitable because
a very small change in sampling or analysis can make a
big difference in final concentrations.
Based on the modeling and measuring results, it ap-
pears that the average atmospheric PHg concentration at
Bahman Square and Bagh Ferdows station is about
0.2 ng/m3
and 0.1 ng/m3
, respectively. However, only
20 measured data during fall and winter appear to be
inadequate to reach solid conclusion; therefore, it is rec-
ommended that additional measurements of mercury for
the entire year are made with a larger number of stations.
It is possible that the differences of PHg concentrations in
modeled and measured days may be due to weaknesses in
WRF simulation of stable conditions especially in winter
season. This issue can be explored in the future works.
Concentrations and dry depositions of different forms
of atmospheric mercury at Bahman Square station are
higher than Bagh Ferdows station due to proximity to
mercury stationary emission sources. Wet and dry depo-
sition of GOM is more than other forms of atmospheric
mercury due to its reactivity. It is required to have mea-
sured data for GOM and GEM concentrations and wet
and dry deposition to determine the accuracy of model-
ing results.
0
10
20
30
40
50
60
70
80
2010-11-08
2010-11-13
2010-11-14
2010-11-15
2010-11-28
2010-11-29
2010-12-05
2010-12-06
2011-01-31
2011-02-06
2011-02-14
d
n
a
M
E
G
,
M
O
G
f
o
n
o
i
t
i
s
o
p
e
D
y
r
D
y
l
i
a
D
PHg
(ng/m
2
)
PHg GOM GEM
(a)
0
20
40
60
80
100
120
140
2010-11-08
2010-11-13
2010-11-14
2010-11-15
2010-11-28
2010-11-29
2010-12-05
2010-12-06
2011-01-31
2011-02-06
2011-02-14
d
n
a
M
E
G
,
M
O
G
f
o
n
o
i
t
i
s
o
p
e
D
y
r
D
y
l
i
a
D
PHg
(ng/m
2
)
PHg GOM GEM
(b)
0
20
40
60
80
100
120
140
160
2010-11-08
2010-11-13
2010-11-14
2010-11-15
2010-11-28
2010-11-29
2010-12-05
2010-12-06
2011-01-31
2011-02-06
2011-02-14
m
/
g
n
(
n
o
i
t
i
s
o
p
e
D
y
r
D
l
a
t
o
T
y
l
i
a
D
2
)
Bagh Ferdows Bahman square
(c)
Fig. 10 Modeled daily dry deposition of GOM, GEM, and PHg in Bagh Ferdows station (a), Bahman Square station (b), and daily total dry deposition (c)
Air Qual Atmos Health (2020) 13:529–541 539

12. Fig. 11 Wet deposition of PHg (a), GEM (b), and GOM (c) in the total domain for 31.01.2011 at 06:00 UTC. The boundaries of Tehran city were used
0
0.02
0.04
0.06
0.08
0.1
0.12
0
20
40
60
80
100
120
140
1 3 5 7 9 11 13 15 17 19 21 23
Hourly
Wet
Deposion
of
GEM
(ng/m2)
g
H
P
d
n
a
M
O
G
f
o
n
o
i
t
i
s
o
p
e
D
t
e
W
y
l
r
u
o
H
(ng/m2)
Hours
PHg GOM GEM
(a)
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0
20
40
60
80
100
1 3 5 7 9 11 13 15 17 19 21 23
Hourly
Wet
Deposion
of
GEM
(ng/m2)
g
H
P
d
n
a
M
O
G
f
o
n
o
i
t
i
s
o
p
e
D
t
e
W
y
l
r
u
o
H
(ng/m2)
Hours
PHg GOM GEM
(b)
0
20
40
60
80
100
120
140
160
180
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
m
/
g
n
(
n
o
i
t
i
s
o
p
e
D
t
e
W
l
a
t
o
T
y
l
r
u
o
H
2)
Hours
Bagh Ferdows Bahman square
(c)
Fig. 12 Modeled hourly wet deposition of GOM, GEM, and PHg in Bagh Ferdows station (a), Bahman Square station (b), and hourly total wet
deposition (c)
Air Qual Atmos Health (2020) 13:529–541
540

13. Most previous studies of atmospheric mercury simulations with
the CMAQ model have been conducted at regional scales. This
study shows that CMAQ alsohas the abilitytosimulatemercuryat
the urban scales. However, many of the effective parameters re-
garding the fate of mercury in the atmosphere and its reactions,
especially in reactive gaseous form, are still unknown.
Acknowledgments The authors would like to thank the Tehran Air Quality
Control Company (AQCC) for providing the atmospheric particulate mercury
measuring data in Bagh Ferdows and Bahman Square stations.
References
Air Quality Control Company (2011) Heavy Metals Measurements in Air
of Tehran City. AQCC, Tehran
Bullock O (2000) Modeling assessment of transport and deposition pat-
terns of anthropogenic mercury air emissions in the United States
and Canada. Sci Total Environ 259:145–157. https://doi.org/10.
1016/s0048-9697(00)00578-7
Bullock O, Brehme K (2002) Atmospheric mercury simulation using the
CMAQ model: formulation description and analysis of wet deposi-
tion results. Atmos Environ 36:2135–2146. https://doi.org/10.1016/
S1352-2310(02)00220-0
Driscoll C, Mason R, Chan H, Jacob D, Pirrone N (2013) Mercury as a
global pollutant: sources, pathways, and effects. Environ Sci
Technol 47:4967–4983. https://doi.org/10.1021/es305071v
EPA (1994) Emission factor documentation for AP-42, Section 11.6,
Portland Cement Manufacturing, Final Report for U. S.
Environmental Protection Agency. Office of Air Quality Planning
and Standards, Emission Inventory Branch. Available from https://
www3.epa.gov/ttn/chief/ap42/ch11/final/c11s06.pdf
EPA (1997) Emission factor documentation for AP-42, section 2.4,
Municipal solid waste landfills, revised. Office of Air Quality
Planning and Standards Office of Air and Radiation, U.S.
Environmental Protection Agency. Available from https://www3.
epa.gov/ttn/chief/ap42/ch02/final/c02s04.pdf
EPA (1998a) Emission factor documentation for AP-42 section 1.4
Natural gas combustion. Eastern Research Group, Morrisville,
North Carolina. Available from https://www3.epa.gov/ttn/chief/
ap42/ch01/final/c01s04.pdf
EPA (1998b) Report on revisions to 5th edition AP-42 Section 1.3 Fuel
Oil Combustion. Eastern Research Group, Morrisville, North
Carolina. Available from https://www3.epa.gov/ttn/chief/ap42/
ch01/final/c01s03.pdf
EPA (2015) Operational Guidance for the Community Multiscale Air
Quality (CMAQ) Modeling System Version 5.1. Prepared in coop-
eration with the Community Modeling and Analysis System
Institute for the Environment University of North Carolina at
Chapel Hill Chapel Hill, NC. Available from https://www.
airqualitymodeling.org/index.php/CMAQ_version_5.1_
(November_2015_release)_Technical_Documentation
EPA (2013) SMOKE v3.5.1 User’s Manual. The institute for the
Environment, The University of North Carolina at Chapel Hill.
Available from https://www.cmascenter.org/smoke/documentation/
3.5.1/manual_smokev351.pdf
Gbor P, Wen D, Meng F, Yang F, Sloan J (2007) Modeling of mercury
emission, transport and deposition in North America. Atmos
Environ 41:1135–1149. https://doi.org/10.1016/j.atmosenv.2006.
10.005
Ghader S, Yazgi D, Shahbazi H (2016) Prediction of wind and other
meteorological fields over Tehran area using an ensemble forecast-
ing system developed for WRF model to provide inputs of air qual-
ity models. 4th national conference on air and noise pollution man-
agement (AQM2016), Tehran
Holloway T, Voigt C, Morton J, Spak S, Rutter A, Schauer J (2012) An
assessment of atmospheric mercury in the Community Multiscale
Air Quality (CMAQ) model at an urban site and a rural site in the
Great Lakes Region of North America. Atmos Chem Phys 12:7117–
7133. https://doi.org/10.5194/acp-12-7117-2012
Marrugo-Negrete J, Urango-Cardenas I, Nunez S, Diez S (2014)
Atmospheric deposition of heavy metals in the mining area of the
San Jorge river basin, Colombia. Air Qual Atmos Health 7:577–588.
https://doi.org/10.1007/s11869-014-0260-0
National Center for Atmospheric Research (2017) ARW Version 3
Modeling System User’s Guide, Mesoscale Microscale
Meteorology Division. Available from https://www2.mmm.ucar.
edu/wrf/users/docs/user_guide_V3.9/contents.html
Rice K, Walker E, Wu M, Gillette C, Blough E (2014) Environmental
mercury and its toxic effects. J Prev Med Public Health 47:74–83.
https://doi.org/10.3961/jpmph.2014.47.2.74
Ryaboshapko A, Bullock O, Christensen J, Cohen M, Dastoor A, Ilyin I,
Petersen G, Syrakov D, Artz R, Davignon D, Draxler R, Munthe J
(2007a) Intercomparison study of atmospheric mercury models: 1.
Comparison of models with short-term measurements. Sci Total
Environ 376:228–240. https://doi.org/10.1016/j.scitotenv.2007.01.
072
Ryaboshapko A, Bullock O, Christensen J, Cohen M, Dastoor A, Ilyin I,
Petersen G, Syrakov D, Travnikov O, Artz R, Davignon D, Draxler
R, Munthe J, Pacyna J (2007b) Intercomparison study of atmospher-
ic mercury models: 2. Modelling study vs long-term observations
and comparison of country deposition budgets. Sci Total Environ
377:319–333. https://doi.org/10.1016/j.scitotenv.2007.01.071
Shahrokhi N (2015) Determination of sources and amount of mercury
emitted from human activities in Tehran City. Msc. theses,
University of Tehran, Tehran
Travnikov O (2005) Contribution of the intercontinental atmospheric
transport to mercury pollution in the Northern Hemisphere. Atmos
Environ 39:7541–7548. https://doi.org/10.1016/j.atmosenv.2005.
07.066
UNEP (2017) Toolkit for identification and quantification of mercury
releases, Reference Report and Guideline for Inventory Level 2,
Version 1.4. Available from https://wedocs.unep.org/bitstream/
handle/20.500.11822/14781/UNEP-Hg-Toolkit-Reference-Report-
January2017.pdf?sequence=1isAllowed=y
Walcek C, De Santis S, Gentile T (2003) Preparation of mercury emis-
sions inventory for eastern North America. Environ Pollut 123:375–
381. https://doi.org/10.1016/S0269-7491(03)00028-9
Wip D, Warneke T, Petersen A, Notholt J, Temme C, Kock H, Cordy P
(2013) Urban mercury pollution in the City of Paramaribo,
Suriname. Air Qual Atmos Health 6:205–213. https://doi.org/10.
1007/s11869-011-0162-3
Zhou H, Hopke P, Zhou C, Holsen T (2019) Ambient mercury source
identification at a New York State urban site: Rochester, NY. Sci
Total Environ 650:1327–1337. https://doi.org/10.1016/j.scitotenv.
2018.09.040
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.
Air Qual Atmos Health (2020) 13:529–541 541