vvEPA
EPA/600/R-09/114 I September 2009 | vwvw.epa.gov/athens
United States
Environmental Protection
Agency
Development and Example
Application of a Pilot Model for
the Biogeochemical Cycling of
Mercury in Watersheds:
SERAFM-NPS
Ecosystems Research Division, Athens, GA 30605
National Exposure Research Laboratory
Office of Research and Development
-------
EPA/600/R-09/114
September 2009
Development and Example
Application of a Pilot Model for the
Biogeochemical Cycling of Mercury in
Watersheds: SERAFM-NPS
by
Christopher D. Knightes
Ecosystems Research Division
National Exposure Research Laboratory
Athens, GA
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
Athens, GA 30605
-------
Abstract
Mercury is a developmental neurotoxicant, ubiquitous in the environment, existing both
naturally and through anthropogenic additions, resulting in human and ecological
exposure risks primarily via consumption of mercury contaminated fish tissue. To better
understand the risk associated with mercury exposure, it has become necessary to not
only understand the mercury biogeochemical cycling within water bodies where typical
mercury exposure occurs, but to also understand terrestrial mercury biogeochemical
cycling, including mercury deposition, transformation, and transport to receiving water
bodies. Here, we present a relatively straight-forward and transparent spreadsheet-based
pilot model to simulate the biogeochemical cycling of mercury in watersheds. The
watershed is divided into different land use types (currently impervious, forest, grassland,
agriculture-pasture, agriculture-row crops, and wetlands) lumping all similar land use
types into one box. This model uses a simple box-model approach, with mechanistic
differential mass balance equations to describe the transformation and transport of
speciated mercury (Hg(0), Hg(II), and MeHg) within each land use type, predicting soil
mercury concentrations and transport processes (volatilization, erosion, leaching, runoff,
and total flux to receiving water bodies). The model is dynamic, running on time steps of
years, allowing for development of mercury concentrations over long time periods. The
output of this model was designed to provide loading information to water body models
such as SERAFM and WASP.
11
-------
Table of Contents
Abstract ii
Table of Contents iii
1 Introduction
2. Model Overview
3. Model Setup
31 Lan d use Cha.ract.eri /ati on
3 2
3.3.
3.4.
3 5
3.5.
Land use Characterization
Mercury Deposition to Watershed
Transport Processes
Erosion
. 1 . Importance and Model Sensitivity to Erosion
1
O
o
3
3
4
5
6
6
6
3.6. Mercury Cycling Calculations for Each Land use Type 7
4. Application of SERAFM-NPS at Eagle Butte/Lee Dam, SD 9
REFERENCES 10
in
-------
1. Introduction
Mercury is a global pollutant that is ubiquitous in the environment. In the atmosphere,
mercury occurs primarily in its neutral, elemental state (Hg°, Hg(0)), while in the
terrestrial soils, water, and sediments it primarily occurs in its oxidized, divalent state
(Hg2+, Hg(II)) (Morel et al., 1998). Divalent mercury can also be transformed into the
environmentally relevant organic form, methylmercury (CHaHg+, MeHg). The USEPA,
the United States Food and Drug Administration (US FDA), and the European Food
Safety Agency (EFSA) have recognized that methylmercury is a contaminant of concern
in announcing consumer advisories for methylmercury concentrations in fish (USDHHS
and USEPA, 2004; EFSA, 2004). MeHg exposure causes severe human health effects
including immune system suppression, neurodevelopmental delays in children, and
compromised cardiovascular health in adults (Mergler et al., 2007). National human
health data from 1999 - 2002 suggest that 300,000 - 600,000 children are born each year
with blood mercury levels that exceed the U.S. EPA's reference dose for MeHg
(Mahaffey et al., 2004, Trasande et al., 2005).
Methylmercury bioaccumulates (i.e., increases in concentration in an organism during its
period of exposure) and biomagnifies (i.e., increases in concentration from trophic level
to trophic level (e.g., from phytoplankton to zooplankton, to prey fish, to predator fish))
within a given food web. Methylmercury concentrations in fish and piscivorous wildlife
can be a million times more than the aqueous methylmercury concentrations in surface
waters (Jackson, 1998). The ingestion offish tissue contaminated with methylmercury is
the predominant exposure pathway for humans and wildlife. Wildlife exposure to
mercury can be of greater concern than humans because wildlife may survive solely by
eating aquatic organisms, and the management strategy of issuing fish advisories to
specific water bodies cannot reduce wildlife ingestion of contaminated fish. The
2005/2006 National Listing of Fish Advisories (NLFA) by the USEPA reported that there
are 3,080 advisories for mercury in 48 states, 1 territory, and 2 tribes in 2006, up from
2,682 in 2005 and 2,436 in 2004. These advisories represent a total of 14,177,175 lake
acres and 882,963 river miles. (USEPA, 2007).
Mercury cycling is complex, requiring a multi-component understanding of mercury,
which encompasses the three dominant mercury species: Hg(0), Hg(II), and MeHg.
Mercury is particularly challenging because of the dominance of different mercury
species depending on the media of interest. Mercury first enters the global cycle through
both anthropogenic and natural sources. Anthropogenic point sources of mercury consist
of combustion (e.g., utility boilers, municipal waste combustors, commercial/industrial
boilers, medical waste incinerators) and manufacturing sources (e.g., chlor-alkali,
cement, pulp and paper manufacturing) (USEPA, 1997). Natural sources of mercury arise
from geothermic emissions such as crustal degassing in the deep ocean and volcanoes as
well as dissolution of mercury from geologic sources (Rasmussen, 1994).
Over the past decade, the Ecosystems Research Division of the National Exposure
Research Laboratory of the Office of Research and Development has been involved with
-------
mercury exposure modeling research. This research, as with most previous mercury
research and management efforts, has focused on mercury cycling in water bodies and
consequent bioaccumulation of mercury in fish and wildlife. More recent studies of
smaller lake (Knightes and Ambrose, 2007) and riverine systems (Knightes et al., 2009,
Brigham et al., 2009, Marvin-DiPasquale et al., 2009) have demonstrated the need for
understanding the mercury cycling within watersheds, and the release of mercury from
watersheds to receiving water bodies, to fully understand mercury exposure in aquatic
ecosystems.
Watersheds receive and transform atmospheric deposition of mercury and transport
mercury to their associated water bodies (e.g., lakes or rivers). Watershed mass balance
work suggests that 10 to 20 % of the deposited mercury is transported through the
watershed (Rudd, 1997), with the other 80 to 90 % either lost to ground water through
leaching, returned to the atmosphere via volatilization, or stored within the terrestrial
surface. Recent work with the METAALICUS study in Canada has suggested that much
of the mercury deposition (from events?) over several recent years remains in the
watershed, and very little of the recently deposited mercury leaks out to water bodies.
This suggests that the build-up of mercury in watersheds is very slow and that movement
within the terrestrial system is similarly slow, possibly on the decadal scale (Harris et al.,
2007).
To improve the incorporation and importance of watershed mercury cycling in
watersheds and its impact on associated surface water bodies, a pilot model has been
developed. This model, SERAFM-NPS (Spreadsheet-based Ecological Risk Assessment
for the Fate of Mercury - Non-Point Source) is a dynamic differential, mechanistic mass
balance model, which simultaneously solves the governing equations for three mercury
species in different land use compartments. The SERAFM-NPS model was developed in
a spreadsheet environment to present a straight-forward and transparent medium for
mercury calculations. By choosing a spreadsheet framework, the model can be easily
adapted and expanded. A user can create a different representation for a given governing
process in the system, add a new process, or even add an additional worksheet for a
different land use type. The spreadsheet is transparent by allowing anyone to see exactly
how the model is performing calculations; the equations and parameters are readily
available, and all interim calculations are readily apparent. The SERAFM-NPS model
calculates mercury species loading for Hg(II), Hg(0) and MeHg into receiving water
bodies (at the pour point) for each of the modeled land uses and sums these results to
provide total loads for all species. These outputs may then be linked to water body
models as loading functions (such as SERAFM (Knightes, 2008) and WASP (Ambrose et
al., 1993)). The model additionally provides soil mercury concentrations and flux rates
for the modeled transport processes. Details of the modeling approach and formulation
are presented below. Section 2 provides an overview of the model structure and model
outputs. Section 3 describes how the model is set up, with more descriptions of model
processes for each land use and how land uses differ, how the model is used, and the
methodology of the cycling calculations. Section 4 provides an example of a SERAFM-
NPS application.
-------
2. Model Overview
The SERAFM-NPS (Spreadsheet-based Ecological Risk Assessment for the Fate of
Mercury - Non-Point Source) model was designed to simulate the cycling of mercury
within a watershed, including Hg(0), inorganic divalent mercury (Hg(II)), and
methylmercury (MeHg). The model incorporates the transport processes of wet and dry
deposition, evasion, leaching, and runoff; and the transformation processes of reduction,
oxidation, methylation and demethylation. The model is designed to lump all land uses of
a given type into a single box. Therefore, the model does not account for spatial
variability of land use across the watershed being modeled. Each land use type has its
own parameterization based on the given land use characteristics. All mercury cycling in
each land use is modeled by representing the land as a uniform, well-mixed batch reactor.
The amount of mercury flux from each land use is calculated, and the total mercury flux
from the watershed is directed to a common pour point. Currently SERAFM-NPS is
designed to model six different land use types: impervious, forest, grassland, agriculture
(pasture), agriculture (row crops), and wetlands. SERAFM-NPS simulates erosion using
RUSLE (Revised Universal Soil Loss Equation (Renard, et al., 1996)), however runoff is
not modeled directly and relies on user-inputted runoff. Figure 1 shows how SERAFM-
NPS sets up the modeling structure given a watershed, by lumping all the same land use
types into individual, well-mixed reactor that are modeled separately. Then, the outflows
from each of these boxes are summed to determine the total mercury flux for each
mercury species to the common pour point. Figure 2 graphically represents the transport
processes, transformation processes, and the transformation linkages between mercury
species. Each land use, well-mixed reactor is modeled as represented in Figure 2, with
different parameterization dependent on the given land use type.
3. Model Setup
3.1. Land use Characterization
The SERAFM-NPS model is written in a spreadsheet format (Microsoft ® Office Excel,
2003) and presented using each worksheet as an effective model program subroutine. The
first worksheet is "Watershed Input File," which is the main input parameterization
worksheet for the model. The user needs to assign percentages for the land use types:
impervious, forest, grassland, agriculture (pasture), agriculture (row crops), and wetlands.
This can be directly entered in the worksheet "Watershed Input File" in cells B4 to BIO.
The watershed area is put into cell B3. If the user has a GIS layer from NLCD, a user can
use a GIS tool to calculate the percentage of each land use type according to the NLCD
classifications (see Table 1) and enter them into the "Watershed Erosion" spreadsheet
along with the watershed area, where the model will then collapse the more refined
classifications into the six land use types required for SERAFM-NPS. These are directly
linked to the "Watershed Input File." By setting up the watershed land use classifications,
SERAFM-NPS divides the watershed area and apportions it to the different land uses.
This effectively creates six boxes with sizes based on the apportioned areas; therefore, the
-------
watershed characteristics and simulations are not spatially explicit. Each lumped land
use type transports a mercury load directly to the shared, single pour point.
3.2. Land Use Characterization
Once the watershed is divided by land use, the next step is to parameterize the
characteristics associated with each land use. This is done within the "Watershed Input
File," with distinction given between each land use type. Each land use has its own
specific column for parameterization. The user should go through each land use type and
enter the appropriate parameter. For two parameters, soil type and organic matter, the
user is limited to the choices given in the drop down menu.
Cells marked with yellow in this worksheet denote cells that the model will populate.
The other cells are color coded to make the model more user friendly. The blue cells are
for the watershed characterization, typically populated via linkage to the land use
characterization in the "Watershed Data" worksheet. The specific land uses are also
color-coded to improve ease of use: grey (urban/impervious), dark green (forest), light
green (grassland), bright green (agriculture - pasture), yellow-orange (agriculture - row
crops), bright blue (wetlands). This color coding is also used for the tabs to visually
assist with finding the tabs associated with each land use (note: shades of green were
used to roughly reflect the land use type but also to avoid using greens and reds because
of color blindness sensitivities).
The cells that are calculated within this worksheet are the fractions of mercury partitioned
between the different phases: air, solids, and water. Hg(0) is modeled to partition
between the air, water, and solids (though currently, the partition coefficient between
water and solids is set to zero, so there is no Hg(0) sorbed to solids) with a fraction of the
total soil concentration in the gas phase, the water phase, and solids phase (though for
Hg(0),/, = 0 currently):
g
RT
H_
RT
TT
~ryrr
Kl
0,.
Equation 1
Equation 2
Equation 3
where:
fg fraction in the gas phase [--]
f-w fraction in the water phase [—]
fs fraction in/on the solid phase [-]
9V soil volumetric void content [cmVcm3]
6-w soil volumetric water content [cm3/cm3]
-------
H Henry ' s Law constant [atm-m3/mole]
R Universal gas constant [atm-m3/mole-K]
T Temperature [K]
Kd partition coefficient between water and soil [L/kg]
Pd bulk density [g/cm3]
Hg(II) and MeHg partitions between the water and solids, with a fraction of the total soil
concentration in the water phase and on the solid phase:
fw = - - Equation 4
fs=l-fw Equations
Mercury transformation rate constants are also entered on the "Watershed Input File" in
units of per day. These units are set up so that they can be different for the different land
use types. These rate constants are set up to be first order and correspond with the
processes depicted in Figure 2. Also entered on the "Watershed Input File" are the
starting soil concentrations for HgT, MeHg, and Hg(II), in rows 54 to 56 for each land
use type.
3.3. Mercury Deposition to Watershed
Atmospheric mercury deposition to the watershed is the primary forcing function driving
mercury soil concentrations and subsequent flux from the watershed to a receiving water
body. The fractions of wet deposition and dry deposition that are each mercury species
are entered in cells B13 - B15 and C13 - C15. The fractions act on the corresponding
total deposition rates, which are entered in the "Hydrology" worksheet.
The "Hydrology" worksheet presents the overall transport fluxes governing mercury
movement. In the first section, the annual precipitation [cm/yr] is entered in column C.
The annual mean air temperature [°C] is entered in column D. Atmospheric Hg(0)
[ng/m3] is entered in column E. Total mercury (HgT [ug/m3]) is entered in column F.
Wet deposition flux [ug/m2/yr] is calculated as precipitation multiplied by the HgT
concentration in rain. Dry deposition flux [ug/m2/yr], column H, defaults to be equal to
wet deposition flux, but can be overridden if more site-specific data is available. Total
deposition flux, column I, is the sum of wet and dry deposition. All of these parameters
are allowed to vary at different time steps for investigation into systems where these
parameters change. Currently the time step is set up in a number of years, as defined by
the user in cell Cl. The default value is 10 yrs. Since the model is set up with Excel, the
number of years can easily be altered and the length of the run can be altered. If the time
step is altered here, it is important to make sure that the parameters that are allowed to
change over time are altered respectively.
-------
3.4. Transport Processes
The "Hydrology" worksheet is the foundation for where the dynamic nature of this model
is initiated. The time step is currently set for 10 yrs, but can easily be altered. Each land
use type has its own section to allow for different transport velocities [m/yr] for each of
the transport processes. The transport processes include: irrigation, evapotranspiration
(ET), leaching and runoff. Currently evapotranspiration and irrigation are not used in the
calculation but are included for completeness and as placeholders in case a water balance
is included in future development. The transport flows of runoff and leaching are used in
the mass balance of mercury in the land use specific tabs.
3.5. Erosion
The parameterization and calculations for erosion for all land use types are performed in
the worksheet "Watershed Erosion." The SERAFM-NPS model explicitly models erosion
using the Revised Universal Soil Loss Equation (RUSLE) as described in detail in
Renard, et al. (1996). The RUSLE calculates the average annual soil loss, A, as kg
soil/m2/yr, as a multiplication of five factors.
0.224-
kg/m2
tons I acre
Equation 6
where: A is average annual soil loss [kg/km2-yr]
R is Rainfall/runoff erosivity factor [kg/km2-yr]
K is Soil erodibility [(tons/acre)/(kg/km2)]
LS is Topographic factor [-]
C is Cover Management factor [—]
P is Support Practice factor [—]
0.224 is Units Conversion [(kg/m2)/(tons/acre)]
Tables including parameters necessary for the RUSLE are organized in the worksheet
"Data Files." Most of the cells in this sheet are linked to the "Watershed Input File."
These cells should not be manually changed on the "Watershed Erosion" sheet because
this will create a broken link between where the data is used and where most parameters
are compiled. This could cause the user to think one number is being used while another
is presented in the "Watershed Input File." Other cells are calculated using the linked
cells. However, R, the rainfall/runoff erosivity factor is location dependent. Therefore, R
is set up as a user input parameter. Renard, et al. (1996) provide a series of maps and
possible adjustments for R.
3.5.1. Importance and Model Sensitivity to Erosion
-------
Because MeHg and Hg(II) sorb strongly to soil particles, erosion is an important transport
mechanism for mercury leaving the watershed. The rate of erosion can great affect the
mercury flux from the given land use type as well as the accumulation and concentration
of mercury in the soils. Therefore, a good understanding and representation of erosion for
each land use type is necessary. Since the soil loss is the product of several factors, an
order of magnitude change in any parameter results in corresponding order of magnitude
change in soil loss.
3.6. Mercury Cycling Calculations for Each Land use Type
Each land use type has its own worksheet where mercury cycling is modeled and mercury
concentrations and mercury fluxes are calculated. The worksheets are: "Urban Hg,"
"Forest Hg," "Grassland Hg," ' Ag Pasture Hg," "Ag Row Crop Hg," and "Wetlands Hg."
The mercury soil concentrations and mercury fluxes are calculated for each of the
separate land use types. The structure of each worksheet is the same for each of the
systems, with minor differences across the different land use types.
All cells on these worksheets are either linked to another worksheet or calculated using
other cells. Therefore, none of the cells on this sheet should be changed. If they are
edited, then links will be broken that are internal to the spreadsheet calculations.
Different sections of the worksheets handle the different processes and calculations. The
first section presents the mercury soil concentration over the time of simulation and the
fraction of that soil concentration that is MeHg and Hg(0). The next section links and
calculates the transport flow rates (runoff and leaching, m3/yr] and solid loads [g/yr].
The next section is the total mercury flux, predicting all the fluxes for mercury in each
land use. The next three sections calculate the mercury concentrations in soils, the
sources and sinks, and the total loading [g/yr] and normalized areal flux [ug/m2-yr] to
water.
The overall equation governing soil mercury concentrations is:
T.-"r
ph x V x —- = Loadi - Lossi + Source. - Sinki Equation 7
dt
Where V is Soil Volume [m3]
Pb is Bulk density [kgsoii/m3]
Ct is Concentration of species I [ugHg/kgSoii]
/ is Species (Hg(0), Hg(II), MeHg)
t is time [yr]
Loadi is Load added to system [gng/yr]
is Loss removed from system [gng /yr]
is Source added to system [gHg /yr]
is Sink removed from system [gng /yr]
-------
This equation is solved for Hg(II) and MeHg for all land use types by using forward
Euler finite difference as shown:
_.,, _. (Load '.- Loss .+ Source - Sink ) . _ .
C\+l = C\ + ± - '• - '• - '• - lj- x At Equation 8
where
C/+1 is Concentration of species / at the time t+1 [ugHg/kgsoii]
C\ is Concentration of species /' at the time t [ugHg/kgsoii]
Pb is Bulk density [kgsoii/m3]
The runoff is divided into impervious runoff flow and pervious runoff flow (m3/yr) based
on the impervious fraction of the given land use type. Erosion load of solids (g/yr) is
calculated using the RUSLE calculations.
The mercury fluxes (ug/m2/yr) of deposition, runoff, volatilization, erosion and leaching
are calculated next. Deposition is linked from the "Hydrology" worksheet. Runoff,
volatilization, erosion, and leaching are all calculated using the soil concentrations
predicted for each species of mercury on each land use worksheet. The total mercury
fluxes are the sums of each of the processes governing each of the mercury species.
Here, mercury flux is HgT.
The next three sections of the worksheet calculate the individual processes for the
individual mercury species, Hg(II), MeHg, and Hg(0), respectively. Here the mercury soil
concentrations are calculated by adding the loads and sources and subtracting the losses.
Using the phase fractions, the concentrations of each of the mercury species is calculated
in pore water and gas phase and that sorbed to solids. These calculations are important
because the mercury sorbed to solids will be transported via the erosion flux, while the
dissolved will be transported via leaching and runoff, and the gas phase is used for
volatilization. All of the necessary components of equations 7 and 8 are determined for
each species. First the loads of wet deposition and dry deposition are calculated, and then
the losses of runoff, erosion, and leaching are calculated. Next the sources and sinks are
calculated, which are specific to the mercury species (Table 3). Volatilization of Hg(0) is
modeled as both a load and a loss, whichever is greater determines the net direction of
gaseous Hg(0).
Because of the instability in Equation 8 due to the fast flux of Hg(0), Hg(0) was modeled
differently than Hg(II) and MeHg. A steady-state analytical solution was used to predict
Hg(0) in the soils. This equation is:
Total Load + Sources
= - 7-7 - - — N -
(Runoff Flow + Leaching Flow) x Jw'Hg(0} * Pb + (Sinks^V x pb )
I ^ )
Equation 9
-------
All worksheets are calculated similarly, with differentiation in each land use via
parameterization differences. The only additional difference is that an additional loading
source is present for the Forest land use; this is litterfall. Litterfall accounts for the
mercury present in leaves that is added to the soil matrix. Because litterfall is specific to
forest, cell E27 contains the multiplier for litterfall, which multiplies the wet deposition.
The multiplier is set to a default of 1, so that litterfall is exactly equal to the wet
deposition flux.
4. Application of SERAFM-NPS at Eagle Butte/Lee Dam,
South Dakota, USA
Eagle Butte is located in the north/central portion of South Dakota on the Cheyenne River
Sioux Tribal Lands. The site modeled (Lee Dam) is a shallow, well-mixed farm pond
surrounded mainly by grassland and cultivated cropland with some woody wetlands and
pasture with predominant clay loam soils (Fig. 1). The watershed area is 25.6 km2, with
percentages as listed in Table 1 and then collapsed down in Table 2. A time step of 10 yrs
was used for a total model run of 400 yrs. Deposition was held constant at 10 ug/m2/yr
for wet and for dry (20 ug/m2/yr total deposition) for 200 yrs and then cut in half from
200 to 400 yrs (5 ug/m2/yr, 10 ug/m2/yr total deposition). Three example output files are
presented in the worksheet "Output Files." These are: Figure 3. Mercury in Watershed
Soils, Figure 4. Mercury Loading from Watershed Normalized by Area, and Figure 5.
Total Mercury Loading from Watershed by Land use. Figure 3 shows how soil
concentrations build over time, demonstrating how quickly or slowly the mercury
concentration builds as it approaches a steady-state value. In this example, that steady-
state value is not reached within the 200 yr model time frame, before the loading
decreases and mercury soil concentration decreases. Figure 4 shows how the mercury
loading from each land use type varies. Figure 5 incorporates the areas of each land use.
The differences between Figure 4 and Figure 5 demonstrate the importance of not only
the land use type but the total area of that land use in determining total mercury coming
off the watershed.
-------
REFERENCES
Ambrose, R.B., Jr., T.A. Wool, J.L. Martin. 1993. The Water Quality Analysis
Simulation Program, WASPS, Part A: Model Documentation. US EPA,
Environmental Research Laboratory, Athens, GA.
Brigham, M.E. D.A. Wentz, G.R. Aiken, D.P. Krabbenhoft. 2009. Mercury Cycling in
Stream Ecosystems. 1. Water Column Chemistry and Transport. Environmental
Science & Technology. 43(8): 2720-2725.
EFSA. 2004. Press Release, EFSA Provides Risk Assessment on Mercury in Fish:
Precautionary Advice Given to Vulnerable Groups. March 18.
Harris, R.C., Rudd, J.W.M., Amyot, M., Babiarz, C.L., Beaty, K.G, Blanchfield, P.J.,
Bodaly, R.A., Branfireun, B.A., Gilmour, C.C., Graydon, J.A., Heyes, A.,
Hintelmann, H., Hurley, J.P., Kelly, C.A., Krabbenhoft, D.P., Lindberg, S.E.,
Mason, R.P., Paterson, M.J., Podemski, C.L., Robinson, A., Sandilands, K.A.,
Southworth, G.R., St. Louis, V.L., Tate, M.T. 2007. Whole-ecosystem Study
Shows Rapid Fish-Mercury Response to Changes in Mercury Deposition. PNAS.
104(42): 16586-16591.
Knightes, C.D. 2008. Development and Test Application of a Screening-level Mercury
Fate Model and Tool for Evaluating Wildlife Exposure Risk for Surface Waters
with Mercury-Contaminated Sediments (SERAFM). Environmental Modelling &
Software. 23:495-510.
Knightes, C.D. Ambrose, R.B., Jr. 2007. Evaluating Regional Predictive Capacity of a
Process-Based Mercury Exposure Model, Regional - Mercury Cycling Model.
Environment Toxicology & Chemistry. 26(4):807-815
Knightes, C.D. Sunderland, E.M., Barber, M.C., Johnston, J.M., Ambrose, Jr., R.B. 2009.
Application of Ecosystem-Scale Fate and Bioaccumulation Models to Predict Fish
Mercury Response Times to Changes in Atmospheric Deposition. 28(4): 881-893.
Mahaffey, K.R., R.P. Clickner, C.C. Bodurow. 2004. Blood Organic Mercury and
Dietary Mercury Intake: National Health and Nutrition Examination Survey, 1999
and 2000. Environmental Health Perspectives. 112(5):562-570.
Marvin-DiPasquale, M., M.A. Lutz, M.E. Brigham, D.P. Krabbenhoft, G.R. Aiken, W.H.
Orem, B.D. Hall. 2009. Mercury Cycling in Stream Ecosystems. 2. Benthic
Methylmercury Production and Bed Sediment-Pore Water Partitioning.
Environmental Science & Technology. 43(8): 2726-2732.
10
-------
Mergler, D., H.A. Anderson, L.H.M. Chan, K.R. Mahaffey, M. Murray, M. Sakamoto,
A.H. Stern. 2007. Methylmercury Exposure and Health Effects in Humans: A
Worldwide Concern. Ambio. 36(1):3-11.
Morel, F., Kraepiel, A.M.L. and Amyot, M., 1998. The Chemical Cycle and
Bioaccumulation of Mercury. Annual Reviews of Ecological Systems, 29: 543-
566.
Rasmussen, P.E., 1994. Current Methods of Estimating Atmospheric Mercury Fluxes in
Remote Areas. Environmental Science and Technology, 28(13): 2233-2241.
Renard, K.G., G.R. Foster, G.A. Weesies, O.K. McCool, D.C. Yoder, coordinators. 1996.
Predicting Soil Erosion by Water: A Guide to Conservation Planning with the
Revised Universal Soil Loss Equation (RUSLE). U.S. Department of Agriculture,
Agriculture Handbook No. 703, 404 pp.
Rudd, JWM. 1995. Sources of Methyl Mercury to Freshwater Ecosystems: A Review.
Water, Air, and Soil Pollution. 80: 697-713.
Trasande, L., PJ. Landrigan, C. Schecter. 2005. Public Health and Economic
Consequences of Methyl Mercury Toxicity to the Developing Brain.
Environmental Health Perspectives. 113(5):590-596.
USDHHS and USEPA. 2004. Backgrounder for the 2004 FDA/EPA Consumer Advisory:
What you need to know about mercury in fish and shellfish. EPA-823-F-04-008.
March.
USEPA. 1997. Mercury Study Report to Congress. EPA-452/R-97-005, Office of Air
Quality Planning and Standards, available at: www.epa.gov/mercury/report.htm
USEPA. 2007. Fact Sheet: 2005/2006 National Listing of Fish Advisories. Office of
Water. EPA-823-F-07-003. July. Available at:
www.epa.gov/waterscience/fishadvisories
11
-------
Table 1. Translating NLCD land use description into SERAFM-NPS land use
breakdown.
Land Use Classification
% Land Use Type SERAFM-NPS Classification
Open Water
Perennial Ice/Snow
Low Intensity Residential
High Intensity Residential
Commercial/lndustrial/Transportation
Bare Rock/Sand/Clay
Quarries/Strip Mines, Gravel Pits
Transitional
Deciduous Forest
Evergreen Forest
Mixed Forest
Shrubland
Orchards/Vineyards
Grasslands/Herbaceous
Pasture/Hay
Row Crops
Small Grains
Fallow
Urban/Recreational Grasses
Woody Wetlands
Emergent Herbaceous Wetlands
1.7%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.2%
0.0%
0.0%
0.3%
0.0%
64.7%
4.8%
22.1%
0.0%
0.0%
4.2%
0.4%
0.7%
Open Water
Impervious
Impervious
Impervious
Impervious
Impervious
Impervious
Impervious
Forest
Forest
Forest
Grassland
Grassland
Grassland
Agriculture, Pasture
Agriculture, Row Crops
Agriculture, Row Crops
Agriculture, Row Crops
Grasslands
Wetlands
Wetlands
Table 2. NLCD fully collapsed to SERAFM-NPS Classification
SERAFM-NPS Land Class %
Impervious 0.9%
Forest 0.2%
Grassland 69.2%
Agriculture, Pasture 4.8%
Agriculture, Row Crops 22.1%
Wetlands 1.1%
Table 3. Sources and Sinks for Modeled Mercury Species
Mercury Species
Sources
Sinks
Hg(H)
MeHg
Hg(0)
Oxidation
Demethylation
Methylation
Reduction
Reductive Demethylation
Reduction
Methylation
Reductive Demethylation
Oxidation
12
-------
Figure 1.
Lumping spatially
resolve
heterogeneous
watershed into
spatially
independent
mixed boxes
based on % land
use.
% Impervious
% Forest
% Grassland
%
Agriculture,
Open Water 2%
Impervious 1%
Forest <1%
Grassland 69%
Agriculture,
Pasture 5%
Agriculture,
Row Crops 22%
Wetlands 1%
% Agriculture,
Row Crops
% Wetlands
Total Mercury Species'! Flux to Common Pour Point
Represents Hg
cycling for each
land-use
compartment
-------
Figure 2. Governing mercury transport and transformation processes within each lumped
sub-basin, parameterization and spatial area defined by land-use type.
Deposition (Hg(ll), MeHg)
Volatilization (Hg(0))
reduction
Hg(H) -
Hg(0)
oxidation
demethylation
methylation
reductive demethylation
MeHg
Leaching Mg(ll), Hg(0), MeHg)
Erosion (Hg(ll), MeHg)
Runoff (Hg(ll), Hg(0), MeHg)
-------
0
Mercury in Watershed Soils
Forest
Grassland
Pasture
Row Crop
Urban
50
100 150 200 250
Year
300
350
400
450
Figure 3. Mercury concentrations in the soils of different land-use types in the
watershed example of Eagle Butte/Lee Dam, SD
-------
Total Mercury Loading from Watershed by Land-Use
25.0
Forest
Grassland
Pasture
Row Crop
Urban
0 50 100 150 200 250 300 350 400 450
Year
Figure 4. Mercury loading normalized by area for the watershed example of Eagle
Butte/Lee Dam, SD
-------
Mercury Loading from Watershed Normalized by Area
^
CM
F
"Bi
3
T3
ro
0
_i
1 U.U
9rv
.u -•
8n
.U
7 n
/ .u
6n
.u
5n
.u
A n
't.U
3n
.u
2n
.u
-\ n (
1 .U
n n
i
i
i
^ —
— Forest
Grassland
— Pasture
Row Crop
• uroan
0
50 100 150 200 250
Year
300
350
400
450
Figure 5. Total mercury loading for the watershed example of Eagle Butte/Lee Dam, SD
-------
vvEPA
United States
Environmental Protection
Agency
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-09/114
September 2009
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE
D; detach, or copy this cover, and return to the address in
the upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |