Dispersion Modeling of Mercury Emissions from Coal-Fired Power Plants at Coshocton

and Manchester, Ohio

SANG-SUP LEE and TIM C. KEENER'. Department ot Civil and Environmental Engineering, University ot Cincinnati, Cincinnati, OH

ABSTRACT. Mercury emissions from coal-fired power plants are estimated to contribute to approximately 46% of the total U. S.
anthropogenic mercury emissions and required to be regulated by maximum achievable control technology (MACT) .standards.
Dispersion modeling of mercury emissions using the AERMOD model and the Industrial Source Complex Short Term (ISCST3)
model was conducted for two representative coal-fired power plants at Coshocton and Manchester, Ohio. Atmospheric mercury
concentrations, dry mercury deposition rates, and wet mercury deposition rates were predicted in a 5 x $ km area surrounding
theConesvilleandJM Stuart coal-fred power plants. In addition, the analysis results of meteorological parameters showed that
wet mercury deposition is dependent on precipitation, but dry mercury deposition is influenced by various meteorological factors.

OHIOJ SCI I08(4):5-69, 2008

INTRODUCTION pollutants. AERMOD has been promulgated by EPAasapreferred

Due to the impacts of mercury on environment and human
health, mercury has been assigned as one of the hazardous air
pollutants (HAPs) in the 1990 Clean Air Act Amendments
(Kil^roe .ind others 2002). In additinn, in February 2008, the U.S.
district court required Environmental Protection Agency (EPA)
to regulate mercury emissions under section 112 of the Clean Air
Act which requires the application of the maximum achievable
control technology (MACT) (United States Court, 2008). Of
several anthropogenic mercury emission sources, coal-fired boilers
account for approximately 46% anthropogenic mercury emissions
in the United States (Keating and others 1997). Three forms ot
mercuryexistincoalcombustionflucgasstreamsielemcntal mercury
(Hg"). oxidized mercury (Hg^O- and particulate mercury (Hgi').
The atmospheric mercury is deposited on the soil and waterbody
in the absence ot precipitation (dry deposition) and by cloud
niicrophysics and precipitation (wet deposition). While most ot
Hg^' and Hgi" are known to be deposited on a local and regional
scale, Hg" may traasport globally due to its insolubility to water
(Rice and others 1997).
According to the National Energy Technology Laboratory
(NETL)'s 2007 Coal Power Plant Database {http://www.netl.
doe.gov/energy-anaiyses/technology.html). 33 coal-fired power
plants have being operated in Ohio. Dispersion modeling was
conducted for the top two mercury-emitting coal-fired power
plants. Atmospheric mercury concentrations, dry mercury
deposition rates, and wet mercury deposition rates were predicted
on a 500 m Cartesian grid up to 5 km iar trom those power plants.
Meteorological data used in the dispersion model were analyzed to
examine meteorological influences on mercury deposition rates.
MATERIALS AND METHODS
The AERMOD model and the Industrial Source Complex
Short Term (ISCST3) model were used for dispersion modeling
ofmcrcuryemissionsiromthetoptwomercury-cmittingcoal-fired
power plants in Ohio in 2005. AERMOD can be used to model
the local atmospheric dispersion of mercury, while photochemical
models such as community multi-scale air quality (CMAQ) are
typically used to simulate long-range transport and deposition of
'Corrtspoiiiiinf; author: Tim C. Keener. Departmenr of Civil and Environmen-
tal Engineering, Ujiivcrsiry of Cincinnati, Cintinnaci, OH 45221. Email: Tim.
air dispersion model to replace the ISCST3. Dispersion modeling
was conducted by botb AERMOD and 1SCST3 model systems,
and the results were compared in this study. The features of stack
flue gases and mercury emi.ssion rates of those power plants were
obtained from the NETL's 2007 Coal Power Plant Databa.se and
EPA's Toxics Release Inventory (TRI) database (http://www.
cpa.gov/tricxplorer). respectively, and summarized in Table
1. In addition, AERMOD and ISCST3 model parameters are
summarized inTable2. Meteorolgica! data were obtained from the
meteorological resourceccntcr (http://www.wcbmct.com). Due to
limitations on availability, 1990 meteorological data were used for
both AERMOD and ISCST3 modeling. Considering proximity
to each coal-fired power plant, Columbus meteorological data
were applied to the Concsvillc coal-fired power plant, and Dayton
meteorological data were applied to the f M Stuart coal-fired power
plant. As found in the EPA report to Congress (Rice and others
1997), the mereury species in all the stackfluegases were assumed
to consi.st of 58% Hg", 40% H^\ and 2% Hgi\ The atmospheric
oxidized mercury is expected to be deposited more readily than
the atmospheric elemental mercury. Several parameters such as
difl^usivity and Henry's law coefficient are used to apply a diflcTcnt
deposition rate for different mercury species to the AERMOD
model. The level ot deposition rate is also expressed in terms of
scavenging coefficient in the ISCST3 model. Tlie.se parameters
were obtained from the literature (Rice and others 1997, Wesely
and others 2002, Sullivan and others 2003, Turner and Schulze
2007, Douglas and others 2008).
RESULTS
Annual atmospheric mercury concentrations, annual dry
mercury deposition rates, and annual wet mercury deposition
rates were predicted on a 500 m Cartesian grid of ground level
positions in a 5 X 5 km area surrounding each coal fired power
plant using AERMOD and ISCST3. Table 3 summarizes these
annual valuesaveragcd tor the receptor area (5 x 5 km). As shown
in tbe table, AERMOD predicted similar levels of atmospheric
mercury concentrations and dry deposition rates as ISCST3 for
both power plants, but significantly lower wet deposition rates
than ISCST3.
In addition, dry and wet deposition rates were determined for
each month bytheAERMOD model to evaluate monthly variations
and meteorological effects on mercury deposition. As shown
in Eigure 1, both power plants show similar trends of monthly
 DISPERSION AND DEPOSITION OF ATMOSPHERIC MERCURY VOL. 108

Table 1 variations in mercury deposition. Since similar meteorological

conditions were found to be applied to these power plants, this
Description of coal-fired power plants considered in this study
result indicates that the dry and wet mercury deposition rates may
be influenced by meteorological factors. Therefore, the effects of
Coai-fircd Stack Stack Design Exit Mercury Emission meteorological parameters were analyzed with respect to the dry
Power Plant Height Diameter Velocity Rate deposition rate and the wet deposition rate, respectively. Figure 2
(m) (m) (m/sec) (kg/yrl shows correlation of each selected meteorological parameter with
thedry deposition rate ajid the wet deposition rate for each coal-fired
Concsviilc 137 4.3 28 446 power plant. As shown in the figure, while rainfall parameters have
aclose relationship with wet mercury deposition rate, nosignificant
244 7.9 25J
meteorological factor is found for dry mercury deposition.
137 53 112 To further investigate meteorological effects on dry and wet
deposition, the isopleths for atmospheric mercury concentrations,
244 7.9 24.1 dry mercury deposition rates, and wet mercury deposition rates
are illustrated for the Conesville coal-fired power plant during
February in Figures 3, 4 and 5, respectively. As shown in these
JM Stuart 244 5.8 32.3 3S8
figures, while dry deposition has similar isopleths as atmospheric
244 5^ 32J mercury concentration, wet deposition is mainly found in the
center ofthe emission source.
244 5.8 32.3

244 5.8 32.3

DISCUSSION
The average dry and wet deposition rates during each month
were predicted for two representative Ohio coal-fired power plants
using the AERMOD model. While similar trends of monthly
variations in mercury deposition were found between two coal-
Table 2 fired power plants, the dry deposition rate showed a different
Modeling parameters trend of monthly variation from the wet deposition rate. These
results indicate that the dry and wet mercury deposition rates may
AF,R\iOD Site Characteristic Parameters be influenced by meteorological conditions, but have different
meteorological factors each other. As a result ofcorrelation analysis
Parameters Spring. Summer, Fall Winter
with meteorological data, wet mercury deposition was found to
Surface Albedo 0.2 0.6
be dependent on rainfall parameters. Hence the isopleths showed
that wet mercury deposition is centered in the mercury emission
Bo wen Ratio 1 2 source. On the other hand, a critical meteorological factor was not
found for dry mercury deposition. However, the isopleths drawn
Surface Roughness 1 0.5 for dry mercury deposition were consistent with the isopleths for
atmospheric mercury concentrations. Therefore, wet mercury
AERMOD Deposition Parameters
deposition is dependent on precipitation near the emission
source, but dry mercury deposition is related to the dispersion and
Form ot Diffusivity Diffusivity Cuticular Henry s Law
Mercury in Air in Water Resistance Coefficient transport of atmospheric mercury which are influenced by various
(ciir/s) (anVs) (s/cm) (PamVmol) meteorological factors.

Hg 7.23x10' 6.30 X 10^ 1.0x10' 150

HIT' 6.0 10 = .3.256 X 10 ' 1.0 xIO' 6.0 X 10 '

Table 3
Fine Mass Fraction Mean Particle Diameter
Summary of average annual atmospheric concentration, dry deposition,
and wet deposition in a 5 x 5 km area surrounding each coal-fired
0.8 0.4 fim
power plan! predicted by AERMOD and SCS7'3
ISCST3 Deposition Parameters

Coal-fired Power Plant Conesville JM Stuart

Form of Liquid Scavenging Frozen Scavenging Model AERMOD ISCST3 AERMOD ISCST3
Mercury Coefficient Coefficient
(hr/s.mm) (hr/s.mm) Average annual acniosphtrii:
mercury concentration (ng/m') 0.036 0.085 0.014 0.041
Hg" 3.3 X 10" 1.0x10"
Average annual dry mercury
Hg=^ 2.51 lO-* 5.0 X 10 ^
deposition (f/g/m^) 6.25 3.62 4.34 4.4

Hgf 0.68 fxm 7.0x10'' 2.0x10'

Average annual wet mercury
deposition {(ig/nr) fl.-i? 5.01 13.~3
HgP 3.5 [im 2.8 X I0-* ^.Ox 10-
OHIO JOURNAL OF SCIENCE S. S. LEE AND T. C. KEENER

0.3
Average monthly dry deposition for Conesville plant ) 0.2
Average monthly wet deposition for Conesviile plant )
Average monthly dry deposition for JM Stuart plant x O.2
Average monthly wet deposition for JM Stuart plant

O.2

16

-US
o
o.
o 0.1

O
I I I I I I I I I I I \
O 1 2 3 4 5 6 7 8 9 1O 11 12
Month
FIGURE 1. Average monthly dry mercury deposition rates {y:^/n\^, scaled by 0.2) and average monthly wet mercury deposition rates ([ig/m') predicted for the Concwilie
and JM Siuarr coal-fired power plants by the AHRMOD mode! for year 2005.

Conesville coal-fired power plant

1.0

^ 0.5

O-O
Wind speed Calm Temperfatu Rainfall total hours
o
<-> -0.5
Dry deposition
Wet deposition
-1.0

JM Stuart coal-fired power plant

1.O

0.5

0.0
Wind 5pee d Calm riji s Tempe re Rtf finfll i otal Ra hours
o
-0.5
E3 Dry deposition
O Wet deposition
-1.O

FlGURK 2. Correlation of the selected meteorological parameters with dry deposition and wet deposition for the Conesville (cop) and JM Stuart (bottom) coat-firtd
power plants.
68 DISPERSION AND DEPOSITION OF ATMOSPHERIC MERCURY VOL. 108

^DELORME
Data use subtect to license
DeLorme Deiorme Street Atlas USA 2009 1 1 v=, 2 2y,
wwvw deiorme com MN (7 6' W Data Zoom 10-5

FIGURE 3. Isopicths for average monthly atmospheric mt-rtury concentrations (ng/mVmonth) predicted for rht-Conesville coal-tirtd power plane in l-ebriuiry using
AERMOD.

^ DELORME ml
Data use suCiecr to license
e DeLorme Street Atlas USA 2009 1 1'/.- 2 2Vi
n deiorme com MN 7.6' W> Data Zoom 10-5

FH;URE4. Isoplcths for average monthly dry mercury deposition races (^m^/month) predicted for the Conesville coal-fired power plant in February using AERMOD.
OHIO JOURNAL OF SCIENCE S. S. LEE AND T. C. KEENER 69

mt
Cala use subject to lK:ne
O DeLmme DeLorme Street Atlo USA 2OO9 y. 1 iv. 2 2 y.
delorme com MN (7 6' Data Zoom 10-5
FIGURE 5. Isoplcths for average monthly wet mercury deposition rates (ng/mVmonth) prcdicced for the ConcsvUle coal-frcd power plant in February using AERMOD.

Sullivan T M . Lipfert FD, Morris SM. Rcnninger S. 2003. Assessing the mercury
LITERATURE CITED health risks associated with coal-fired power plas: impacts of local
Douglas S.HaneyJ, Wei Y.Myers T. Hud ist hew sky B. 2008. Mercury deposition
depositions. Proceedings of Air Quality IV Conf-crence; 20()i Scp; Arlington
modeling for the Virginia mercury study. Available from htcp;//www.dcq.
(VA). Available from http://204.154.137.l4/teehnologies/coalpower/ewr/
Stiite.va.iis/exporr/sites/dffaiilt/regulations/documents/App_A-VDEQ_
ai r_quality_reseiirch/heal th_effects/pdfs/l'RH-2 AQIV_Sullivan.pdf
Hg^MoJel ing_Report_FINAL_10-20-08.pdf
Turner DB. Schulze RH, 2007, Atmospheric dispersion modeling, Dallas: Trinity
Keating M H . Beauregjrd D, Benjey WG. Driver L, Maxv^cll W H , Peters W D ,
Consultant, Inc. and Air Ik Waste Management A.ssociation. Chapter 9,
Pope A A. 1997, Mercury study reporrto congress volume I I : An inventory
United States Court ot Appeal for the District of ColumbiaCircuit. 2008, Available
ofanthropogenic mercury emissions in the United States. Washington (DC);
from http://pacer.cadc.uscourts.gov/docs/common/opinions/20082/05-
U.S. Environmental Protection Agency, EPA-452/R-97-004.
IO973.pdf
Kilgroe JO. .Scdman CB. Sriva.ava RK. Ryan JV. Lee CW. Tliomtloe SA. 2002.
Wesely ML. Doskey PV. Shannon JD. 2002. Deposition parametcrizations for the
Control (>t mertury emis\lonv from coal-fired electric utility boilers: interim
industrial source complex (1SC3) model. Argonne (IL): Argonnc National
report. Research Triangle Park (NC): U.S. Hnvironmental Protection Agency.
Laboratory. A N L/ER/TR-01/003.
EPA-600/R-01/109.
Rice CE, Ambrose Jr RB. Bullock Jr OR. Swartout J. 1997. Mercury study
report to congress volume I I I : Fate and Transport of Mercury in the
Environment. Washington (DC): U.S. Environmental Protection Agency.
EPA-452/R-97-004.