United States
Environmental Protection
Agency
July 18, 2013
EPA-734-R-1 3-001
Pesticides in Flooded Applications Model (PFAM):
Conceptualization, Development, Evaluation, and User Guide
Dirk F. Young
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, DC 20460
Key Words: Pesticide, Rice, Risk Assessment, Compartment Model
-------
Abstract
The Pesticides in Flooded Applications Model (PFAM) was developed to facilitate risk
assessments for pesticides used in flooded-agriculture applications such as rice paddies and
cranberries bogs. PFAM was designed around the specific parameters that are typically available
for a pesticide risk assessment, thereby simplifying the assessment process by allowing the user
to concentrate on providing only relevant model inputs. The model considers the fate properties
of pesticides and allows for the specifications of common management practices that are
associated with flooded agriculture such as scheduled water releases and refills. It also allows
for natural water level fluctuations resulting from precipitation and evapotranspiration. The
purpose of this document is to describe the concepts used in the model. Because PFAM was
designed for environmental protection regulatory purposes, the quality acceptance criteria
specified that the model estimates should err on the high side of measured data, but it should not
cause undue burden to stakeholders by being overly conservative. As evaluation results show,
PFAM did tend to err on the high side of the data yet provided more realistic estimates than the
currently used methods, which thereby reduced stakeholder burden. Thus, PFAM satisfactorily
performed as a regulatory model. The code for the mathematics of the model is written in
Fortran 95/2003, while the user interface code is in Visual Basic Dot Net. As is important for a
regulatory model, PFAM is nonproprietary, with the model code and documentation being freely
available.
-------
1. Introduction 5
2. PFAM Conceptual Development 6
2.1 Overview of the Processes in PFAM 6
2.1.1 Flood and Overflow Control 7
2.1.2 Plant Growth 8
2.1.3 Chemical Processes 8
2.1.4 Pesticide Applications 23
2.1.5 Degradates 23
2.2 Computations 24
2.2.1 Initial Conditions 24
2.2.2 Analytical Solution for Concentrations 24
2.3 Postprocessing 25
3. Evaluation of PFAM Using Criteria for Regulatory Assessments 27
3.1 Methods 28
3.1.1 Field Studies 28
3.1.2 Chemical Properties 32
3.1.3 Simulations and Comparisons 33
3.1.4 Hypothetical Regulatory Simulations 34
3.2 Results and Discussion 35
3.2.1 Study A: Arkansas, Pendimethalin Simulation Results 35
3.2.2 Study B: Arkansas, Bispyribac Sodium Simulation Results 38
3.2.3 Study C: Texas, Propiconazole Results 40
3.2.4 Study D: Texas, Carbaryl Results 42
3.2.5 Study E: California, Triclopyr Results 44
3.3 Long-Term Regulatory-Type Evaluation 46
3.4 Summary 49
4. User Guidance 50
4.1.1 Background and Purpose 50
4.1.2 Quick Start 50
4.1.3 Menu Items 50
4.1.4 Chemical Tab Sheet 50
4.1.5 Applications Tab Sheet 53
4.1.6 Location Tab Sheet 53
-------
4.1.7 Flood Events Tab Sheet 54
4.1.8 Crop Tab Sheet 54
4.1.9 Physical Tab Sheet 55
4.1.10 Tab Sheet for Output 56
4.1.11 Degradate 1 57
4.1.12 Degradate 2 57
4.1.13 Run Button 57
5. References 58
-------
1. Introduction
Pesticide use on cranberries, rice, and other applications where a pesticide is used in
conjunction with flooding presents unique issues to pesticide risk assessors trying to estimate
relevant environmental concentrations. For these types of uses, assessors need a model with
special flood-handling features to address issues such as whether the pesticide is applied post- or
pre-flood, water levels that vary over the course of the crop, scheduled water releases and refills,
and flow-through washouts. A relevant model for pesticide exposure assessments would
consider these factors as well as the availability of specific fate parameters for a pesticide risk
assessment.
The current U.S. Environmental Protection Agency (USEPA) model for flooded
applications is similar in concept to the equilibrium model suggested by Johnson (1991) and
delivers rough but protective concentration estimates. That model is an equation that determines
the aqueous concentration of a pesticide that is at equilibrium with 10 cm of water and 1 cm of
sediment (USEPA, 2007a). While simple and effective for environmental protection, it does not
take full advantage of available information such as degradation, management practices such as
flooding and draining, or long-term use of a pesticide. Such simple estimates may provide
protective screening-level estimates, but if a pesticide fails the screen, there is no standard model
to provide more pesticide- and application-specific concentrations for use in higher tier risk
assessments.
The new Pesticides in Flooded Agriculture Model (PFAM) described here was designed
specifically for use in a regulatory setting wherein model inputs and processes will correspond to
the data available during a regulatory assessment. In a regulatory assessment, assessors have
available only a few chemical fate parameters, such as those listed in Table 1-1. For this reason,
an appropriate model would balance the complexity of the model with the available inputs (Crout
et al., 2009; Freni et al., 2009; Ranatunga, et al., 2008). PFAM was designed around the specific
chemical parameters given in Table 1-1, which are typically the only parameters available for a
regulatory pesticide assessment. Thus, PFAM is only as complex as the available data allow.
The balance of complexity with available data is consistent with good modeling practices as
specified by the USEPA (USEPA, 2009a).
PFAM borrows heavily from the mathematical formulation of the processes used in the
model EXAMS (Burns, 2000), which is a USEPA standard for modeling pesticide water quality
for non-volume-varying bodies. Note that there are several models described in the literature
that are aimed at determining water quality resulting from aquatic agriculture (e.g., Johnson,
1991; Jeon et al, 2007; Kim et al, 2008; Karpouzas and Etori, 2006; Tournebize et al., 2006;
Yoshinaga, et al, 2004; Watanabe and Takagi, 2000), but none are specifically designed around
the methods used for pesticide risk assessments, non-proprietary, and freely available for public
inspections, as is desirable for USEPA regulatory models (NRC, 2007, USEPA 2009a).
-------
Table 1-1. Typical Relevant Chemical Parameters Available for a
Pesticide Exposure Assessment for Flooded Applications.
Parameter Notes
Sorption Coefficient (Koc)
As typically defined.
Aerobic Metabolism Rate
Only whole system (solid and
aqueous) degradation rate is available.
Typically for 20 to 25°C.
Anaerobic Metabolism Rate
Only whole system (solid and
aqueous) degradation rate is available.
Typically for 20 to 25°C.
Vapor Pressure
Typically for 20 to 25°C.
Solubility
Typically for 20 to 25°C.
Aquatic Photodegradation Rate
Conducted on thin (mm) aqueous
layer with artificial light.
Hydrolysis
Conducted at pH 5, 7, and 9.
2. PFAM Conceptual Development
2.1 Overview of the Processes in PFAM
PFAM is conceptualized in Figure 2-1 and includes both hydrological processes and
chemical processes. The water body depth may change due to precipitation, refill, drainage,
evaporation, and weir-height changes. The model consists of two regions: a water column and a
benthic region. Each individual region is completely mixed and at equilibrium with all phases
within the individual region, and equilibrium within each region follows a linear isotherm. The
two regions are coupled by a first-order mass-transfer process. Chemical transformation
processes (i.e., hydrolysis, bacterial metabolism, photolysis, and sorption) within each region are
formulations that were heavily borrowed from the USEPA EXAMS model (Burns, 2000).
Changes in water body conditions (temperature, water levels, wind speed, etc) and the resulting
changes in degradation rates occur on a daily time step. A daily time step was selected mainly
because of the availability of a large amount of daily meteorological data (Burns et al., 2007) and
the USEPA's historical use of EXAMS on a daily time step.
-------
Direct pesticide application
to littoral region
plant growth
Volatilization Evaporation Precipitation
water input
water column
Potential Overflow
degradation
water column/benthic
mass transfer
leakage
Benthic region
benthic
degradation
leakage
Adjustable
Weir
Depth
Figure 2-1. Pictorial of PFAM showing hydrological and chemical processes.
2.1.1 Flood and Overflow Control
In this conceptualization (Figure 2-1), water is held in a basin behind a weir. Similar
paddy and weir models have been previously created (Yosinaga, et al. 2004, Jeon et al. 2005,
Khepar et al. 2000). The maximum volume of water is controlled by the weir height, which can
move up or down by user control. Users can schedule changes in water level by means of an
external source of solute-free water. If weir height decreases below the current water level, then
the volume of water along with the associated solute above the new weir height is
instantaneously released. The depth of the water column is calculated from daily precipitation,
refill, drainage, leakage, and evaporation. For any day, the water level is calculated as
for 0
-------
for the day is set to dweir, and the excess water is used in the calculations for washout (see below).
The minimum possible water volume is zero, but for practical purposes, it is set to a small value
(e.g., 10~6 m) to prevent numerical difficulties that are associated with calculations involving
infinity and zero.
The computer implementation of the model allows for automation of the refill
requirements. Refill occurs automatically if the water level reaches a user-specified minimum
depth. The subsequent refill adds enough to reach the user-specified fill level. Additionally, the
model can account for those scenarios in which the user needs to have a constant flow through
the water body. In this case, the depth of the water body is maintained at the weir height with
excess water overflowing the weir. The excess water enters into the washout calculations as
described later.
2.1.2 Plant Growth
Plant growth is based on a simple linear increase in areal coverage of the plant, as
described in the following equations:
/-f \ T' -t- T'
p J /?,max T-T T7 e m
\-*m e )
/,=/„.„, Tm
-------
described by two differential equations — namely, a mass balance on the water column region and
a mass balance on the benthic region:
ds ds
+ DOCl L_ _Q
f"sedl , ^'"DOCl , 1 , ~ ^ I
at at at
~ VlVvo£l (2-3)
DOC ~ X?LC1
Sed2
(® + QL XC1 ~ C2
where
ci = aqueous concentration in water column, [kg/ m3]
C2 = aqueous concentration in benthic region, [kg/ m3]
Csed = concentration of suspended sediment in water column = msed i/vi [kg/m3]
CDOC = concentration of DOC in water column = mooc/vi, [kg/m3]
nisedi = mass of suspended sediment in water column, [kg]
niooci = mass of DOC in water column, [kg]
msed2 = mass of suspended sediment in water column, [kg]
Ssedi = sorbed concentration on suspended sediment in water column, [kg/ kg]
SDOCI = sorbed concentration on suspended DOC in water column, [kg/ kg]
ssed2 = sorbed pesticide concentration on benthic sediment, [kg/ kg]
vi = volume of water in region 1 on the specific day, [m3]
V2 = volume of water in region 2, [m3]
Q = volumetric flow rate of water out of water column, [m3/s]
QL = volumetric leakage flow rate, [m3/s]
co = 1st order water-column-to-benthic mass transfer coefficient, [m3/s]
Mhydr = ^ order hydrolysis rate coefficient, [s"1]
Uphoto =lst order photolysis rate coefficient, [s"1]
|j,voi = effective 1st order volatilization rate coefficient, [s"1]
Hbio-ai=lst order aqueous-phase metabolic degradation rate coefficient in water column, [s"1]
= 1st order sediment-sorbed metabolic degradation rate coefficient in water column, [s"1]
= 1st order DOC-sorbed metabolic degradation rate coefficient in water column, [s"1]
Hbio-a2 =lst order aqueous-phase metabolic degradation rate coefficient in benthic region, [s"1]
Hbio-sed2 = 1st order sediment-sorbed metabolic degradation rate coefficient in benthic region, [s"1]
In this model (as well as in the current regulatory use of the EXAMS model) the
following assumptions are made: (1) suspended matter in the water column occupies negligible
volume, (2) hydrolysis, photolysis, and volatilization act only on dissolved species, (3) within a
single region (water column or benthic), the rate coefficient for biological metabolism is the
same for both dissolved and sorbed forms of pesticide (e.g., |j,bioi = Ubio-ai = Ubio-sedi = Mt>io-Doci,
and |j,bio2 = Hbio-a2 = Hbio-sed2), (4) the hydrolysis rate coefficient in the benthic region is the same
as that in the water column, (5) linear isotherm equilibrium exists within each region among all
sorbed species. With these assumptions, we can rewrite equations (2-3) and (2-4) in a simpler
form as follows:
-------
(2-5)
at
tic
^^ = -r2c2+(Q + AXc1-c2) (2-6)
d\.
where T, = — + fwl nphoto + nhydr + n¥0l + -M + nbiol (2-7)
/V,2 (2-8)
(2-9)
Ked2Ksed2+V2)
A = - - ^ - , (2-10)
Ked2Ksed2+V2J
©=
- _
KedlKSedl + mDOClKDOC
where fwi and fW2 are the fractions of solute in the aqueous phase within the water column and
benthic regions, respectively, as defined by
fwi = 1 - ^ - ^ - x (2-12)
KedlKSedl + mDOClKDOCl+Vj
fW2=7 - ^ - x (2-13)
Ked2KSed2 + V2)
and where Ksedi, KDOCI are the linear isotherm partitioning coefficients for suspended sediments,
biota, and DOC in the water column, respectively, and Ksed2 is the linear isotherm partitioning
coefficient for sediment in the benthic region (units of m3/kg).
The term, fwi, for this varying volume model varies on a daily basis depending on the
volume of the water body (vi) as described below in Daily Piecewise Calculations. As a
simplification in this model, the mass of sediment, biota, and DOC remain constant and in
suspension. This assumption has very little impact on the model output in most cases since
partitioning to these species is negligible for all but the most extremely high partitioning
coefficients (described later and in USEPA, 2004).
Given a set of initial conditions, equations (2-5) and (2-6) completely describe the water
body. It is clear, that there are only four parameters that influence the concentration — Fi, F2, Q,
and 0. Fi is the effective overall dissipation rate in the water column region, [s"1]. F2 is the
effective overall degradation rate in the benthic region, [s"1]. Q is a mass transfer coefficient
describing transfer between the benthic region and water column, [s"1]. 0 is the ratio of solute
holding capacity in the benthic region to that in the water column. The following sections
describe the details of these components.
2.1.3.1 Solute-Holding-Capacity Ratio (0)
The solute-holding-capacity ratio (0) is the ratio of solute holding capacity in the benthic
region to the solute capacity in the water column, as defined by equation (2-11). The individual
partitioning coefficients (Ksed and KDOC) in equation (2-11) are generally not directly known for
specific applications. To account for these unknowns, the various partitioning coefficients are
10
-------
related to the organic carbon partitioning coefficient (which is typically known in a pesticide
assessment) by the same relationships used in EXAMS.
For the sediment, the partitioning coefficient is directly proportional to Koc, with the
constant of proportionality equal to the fraction of organic carbon in the sediment. The carbon
amount in the sediment is a user-adjustable input. The sediment partitioning coefficients can thus
be determined from
Ksedl = Ksed2 = focKoco.OO\ (2-14)
where Koc = organic carbon partitioning coefficient, [ml/g]
foc = fraction of organic carbon in sediment [ — ]
Note that the units of the coefficients in equations (2-1) to (2-11) are all given in the s.i form.
The s.i. convention will be maintained throughout this paper. However, for some fundamental
parameters such as Koc, which is usually presented in units of ml/g, the common units will be
used along with the necessary conversion factor.
The partitioning coefficient for DOC is determined from the default empirical
relationships described in the EXAMS documentation (Burns, 2000). PFAM incorporates the
assumption of Burns (2000) that benthic DOC has higher partitioning characteristics than water
column DOC. The relations given by Burns (2000) and adopted for the current and proposed
standard water bodies are as follows:
(2-15)
Figure 2-2 shows an example of the relative capacities of the individual media (aqueous,
DOC, and suspended sediment) in the water column as a function of Koc. With the parameters
from the USEPA standard water bodies for suspended solids and DOC (USEPA, 2004) and with
a 10-cm depth, the water compartment holds 90 percent of the solute up to a Koc value of about
70,000 ml/g. Up to Koc value of about 700,000 ml/g, the aqueous capacity component is greater
than the capacity of all sorbed species in the water column combined
Note that EXAMS and the USEPA standard pond, which were the bases upon which
PFAM was developed, also include a biological partitioning component in the water column.
However, a sensitivity analysis showed that little solute partitioned to the USEPA standard
amount of biological material (0.4 mg/L) except at the highest of Koc values (fraction less than
0.0005 at Koc of 103 ml/g and <0.09 at Koc of 106 ml/g). Furthermore, it is unlikely that
measurements of biological material would be available or would significantly contribute to a
better estimate of pesticide concentrations. Therefore, a biological partitioning component was
not included in PFAM. This elimination is in keeping with the PFAM development ideal to stay
away from unreasonable complexity. Also note, the effect of suspended solids is equally
insignificant; however, the suspended solids perform an additional function in photolysis
quenching, so the suspended solids parameter is retained.
Figure 2-3 shows an example of the relative solute holding capacities for the benthic
region of a typically parameterized water body. During PFAM development, sensitivity analyses
showed that some parameters (i.e., benthic DOC and benthic biota) used in the USEPA EXAMS
pond model (USEPA, 2004) had insignificant impact on results and were thus not included in
PFAM. The only components that had significant impact were the sediment and the pore water
and these are shown in Figure 2-3.
11
-------
A sensitivity analysis was performed on the benthic components used in the USEPA
standard pond, and the relative fractions for the DOC and biota are on the order of 10~4 and 10~6,
respectively, for Koc values of 106 ml/g. For the benthic region, DOC and biota partitioning are
negligible regardless of the Koc value for this parameter set. Therefore benthic biota and benthic
DOC were not included in PFAM.
o
CD
Q.
CD
O
0.8
0.6 --
o
0.4
0.2
o
CD
•Aqueous (10 cm)
- DOC (5 mg/L)
• Suspended Solids (30 mg/L)
10 100 1000 10000 100000 1000000
Koc (ml/g)
Figure 2-2. Relative solute holding capacity of individual components in the water column.
10 100 1000 10000
Koc (ml/g)
100000 1000000
Figure 2-3. Relative solute holding capacity of individual components in the benthic zone. DOC
and biological partitioning fractions are 10 4 or less and are not detectable on this graph.
12
-------
2.1.3.2 Effective Water Column Dissipation (Fi)
The overall dissipation rate in the water column (Fi), as defined in equation (2-7), is the
sum of contributions from hydrologic washout and degradation by mechanisms of biological
metabolism, photolysis, and hydrolysis. The specific methods and assumptions that are used to
determine these individual first-order dissipation processes are described below.
2.1.3.2.1 Hydrologic Washout —
1VJ
The first term in equation (2-7), Q/vi, represents the effective first-order dissipation rate
resulting from flow moving pesticide out of the water body. Flow out of the water body may
occur due to high rainfalls (as dictated by the meteorological input data), by intentional irrigation
flow through, or as leakage through the weir. (Benthic leakage is treated separately in PFAM, see
below). The washout term acts on all forms of pesticide (both aqueous dissolved and sorbed to
suspended matter), as is apparent from equation (2-3). This means that pesticide mass in both
dissolved and suspended sorbed forms can flow out of the water body.
2.1.3.2.2 Water Column Leakage (QL/VI)
The leakage term (Qi/vi) represents the dissipation of the pesticide in the water column
due to leakage of the water column through the benthic region. The assumption here is that only
aqueous-phase pesticide leaks into the sediment and that the leakage rate is constant and only
downward such that there is never leakage in the reverse direction (i.e., into the water column).
Therefore leakage in this conceptualization can only decrease water column concentrations.
Daily leakage volume is constant and occurs until water column is emptied. Further note that
this process is constructed as a first-order process, which facilitates and streamlines the
mathematical formulation and solution methods. Because the depth is assumed constant during
the course of any day, the leakage as a first-order mechanism will be most representative of the
actual process when daily volume changes are small. The assumption will produce more
conservative (protective) results as leakage rate increases and daily depth changes are greater. In
a pesticide risk assessment, the leakage parameter would likely be set to zero, as this would be a
reasonable screening-level approach and would provide conservative estimates for a term that is
difficult to parameterize.
In the registration process of pesticides, an estimate of the aqueous degradation rate under
aerobic conditions is supplied by the registrant. Such estimates are derived from laboratory tests
which are typically conducted in aqueous/sediment systems at 20 to 25°C. These tests generally
cannot differentiate between degradation occurring on the dissolved forms and sorbed forms of
the pesticide; an overall degradation rate is generally all that is determinable from these studies.
Therefore, PFAM treats the sorbed-phase and aqueous-phase degradation rates as the same in the
water column, which makes both equal to the overall rate as described previously under equation
(2-4).
Because temperature impacts degradation rates, an adjustment was included in this
model, which corresponds to the USEPA standard temperature adjustment when data are not
available on temperature effects on metabolism (Burns, 2000). The relationship is as follows:
13
-------
r-biol r-treasured
ilO
10
(2-16)
where (^measured = laboratory measured aerobic metabolism rate , [s"1]
Qio = factor by which degradation increases for a 10°C temperature rise.
T = temperature of modeled water body [°C]
Tref = temperature at which laboratory study was conducted [°C].
In a standard EPA assessment, the QIO is equal to 2, so this temperature modification doubles
the degradation rate for every 10°C rise in temperature. In this model, the water temperature of
the simulations varies on a daily basis. The water temperature is estimated from the backward
30-day average of the daily air temperatures as specified in the meteorological data input.
2.1.3.2.3 Hydrolysis (|ahydr_i)
The hydrolysis degradation acts only on the dissolved phase in the water column. The
hydrolysis rate is directly obtained from experimental measurements, as supplied by pesticide
registrant data submissions. Variations in pH are not explicitly simulated in the model, so the
hydrolysis rate that is used should correspond to the total hydrolysis rate under the conditions
that are to be simulated. It is assumed that hydrolysis acts only on dissolved species. Therefore,
the effective hydrolysis rate is reduced by the factor fwi, as presented in equation (2-7). The
factor fwi represents the fraction of total pesticide that is present in dissolved aqueous form, as
previously described.
2.1.3.2.4 Photolysis
Photolysis rates are derived from standard laboratory tests following EPA-approved
protocols. These tests are designed to estimate the photodegradation rate for near-surface
conditions at specific latitude and under clear-sky conditions. The input value for (J,photo should
be the average value over a 24 hour period. PFAM adopts the methods used in EXAMS (Burns,
2000) to account for latitude adjustments, light attenuation, and cloud cover. These adjustments
are implemented as follows:
M-photo p lat atten Mrneasured \ '
where fp = the fractional area of plant coverage (see equation 2-2)
fiat = latitude adjustment factor, [ — ]
fatten = attenuation factor to absorption, [ — ]
^measured = measured near-surface photolysis rate coefficient at reference latitude and clear
atmospheric conditions [s"1]
The simulated latitude may vary depending on the desired location in the U.S. where a
pesticide assessment is to be made. The effect that latitude has on incident light is accounted for
by the latitude adjustment factor (fiat). This model adopts the latitude adjustment described in the
EXAMS documentation (Burns, 2000). The latitude adjustment is as follows:
_ 191700 + 87050 cos(0. 0349 x Lsim) ,2_lgx
~ 191700 + 87050 cos(0.0349 x Lref )
14
-------
where Lref = reference latitude at which the measured photolysis rate was determined, [degrees]
Lsim = latitude of the simulated scenario, [degrees]
The light attenuation factor (fatten) described by Burns (2000) has also been adopted into
PFAM. Again, full details are given in the EXAMS documentation, and only the resulting
equation is given here:
atten
where Dfac = EXAMS-defmed distribution factor default value = 1.19, [- ]
di = depth of water column, [m]
a = total absorption coefficient, [m"1]
The absorption coefficient (a) is calculated from the EXAMS default conditions — that is,
calculated from the spectral absorption coefficient assuming that the wave length of maximum
absorption occurs at 300 nm. Using the default EXAMS assumptions, the total absorption
coefficient is as follows:
a = 0.141 + 101[CCffi] + 6.25[CDOC] + 0.34[QJ (2-20)
where CDOC, Csed have been previously defined under equation (2-3), and CCHL is the chlorophyll
concentration [mg/L].
As a simplification for this varying-depth model, the concentrations of the physical
components in equation (2-20) remain constant as depth changes. Because this model does not
attempt to simulate the complex sedimentation processes that would inevitably occur with
varying depths, and in keeping with the simple nature of this model, the corresponding
suspended concentration changes were kept constant, with the values for the suspended
concentrations being user inputs. The overall photolysis rate does change, however, due to the
effect of depth on equation (2-19). Figure 2-4 shows a typical expected reduction in the half-life
as a function of depth. When depth is effectively zero (no water volume in the water
compartment) the program switches the photolysis rate to zero. Photolysis on dry soil should be
considered along with the overall dry soil degradation rate.
Temperature affects the photolysis in this model only if the temperature reaches 0°C at
which point photolysis ceases to occur, with the assumption that there will be ice cover below
0°C.
15
-------
20
.-. 18
S
1 16
t!
? 14
o
"o -19
ro '^
LL
10
4
0.05
0.1
0.15
Depth (m)
0.2
0.25
0.3
Figure 2-4. Multiplicative factor for effective half-life in the water column as a function of water
body depth. Measured half-life at 0 degrees latitude; simulated half-life at 34 degrees. Suspended
solids at 30 mg/L, DOC at 5 mg/L; Chlorophyll at 0.005 mg/L. No plant cover.
2.1.3.2.5 Volatilization
The standard water bodies use a two-film model for volatilization calculations, as
described in the EXAMS documentation (Burns, 2000). The concentration of pesticide in the
atmosphere is assumed to be negligible, and thus volatilization becomes a first-order dissipation
process. This model uses all of the volatilization default assumptions described in the EXAMS
documentation. The overall volatilization rate coefficient may be expressed as
(2-21)
where A = surface area of the water column, [m2]
kvoi = volatilization exchange coefficient, [m/s]
The volatilization exchange coefficient is defined in the conventional manner as comprising a
liquid-phase and an air-phase component as follows:
1
1
= -- h
1
(2-22)
where kw = liquid-phase resistance [m/s]
ka = gas-phase resistance, [m/s]
H = Henry's law constant [m3atm/mol]
16
-------
R = the universal gas constant (8.206 x 10~5 m3atm/mol/K)
TK= temperature (K)
This model uses the EXAMS method of referencing the liquid exchange resistance of
pesticides to the liquid resistance of oxygen, and uses molecular weight as a sole surrogate for
molecular diffusivity variations among compounds. Further details can be found in the EXAMS
documentation (Burns 2000). The resulting relationship is as follows:
(2-23)
where ko2 = oxygen exchange constant at 20°C, [m/s]
MW = molecular weight of pesticide.
The oxygen exchange constant is determined from the empirical relationship of Banks
(1975). Adjustments are also made for temperatures other than 20°C. Note that although
EXAMS uses a reference temperature of 20°C for the Banks (1975) relationships, it is not clear
from Banks (1975) what the actual reference temperature should be. Schwarzenbach et al.
(1993), for example, used a 10°C reference for this same relationship. Until this is clarified, the
20°C reference temperature will be used in the model. For wind velocities (vwmd) less than 5.5
m/s, the relationship used is as follows:
k02 =4.19xl(T6>10(l.024^) (2-24)
where wio = wind velocity at 10 m above water surface [m/s].
For wind velocities greater than or equal to 5.5 m/s, the relationship is
km =3.2xl(T7(w10)2(l.024(r~20)) (2-25)
Wind speeds are read from meteorological files in which wind speed is given from
measurements 10 m above the surface (Burns et al., 2007). The following general relation is
used:
UL = log(zi/zo) (2_26)
u2 Iog(z2/z0)
where zo is the boundary roughness height, which is assumed to be 1 mm. For the case where
wind speeds are read from a meteorological file in which wind speed measurements were made
at 10 m, the equivalent wind speed at 0.1 m (uo.i) is as follows:
log(0.1/0.00l) ne
un i = ——, { u, o = 0.5u, o (2-27)
01 Iog(l0/0.00l)
The gas phase resistance is referenced to water vapor resistance, and an empirical
relationship relates the water vapor exchange rate to wind speed. A linear regression of the
17
-------
laboratory-derived data of Liss (1973) is used to develop a correlation to describe the effect of
wind speed on water evaporation rate:
= 0.00005+ 0.0032w,
(U
where ka,H2o = the water vapor exchange velocity (m/s)
uo.i = wind speed velocity measured at 0.1 m above the surface (m/s)
The exchange rate of a pesticide is related to the exchange rate of water by
k. =k.
,H2O
D
D
a,H2O
(2-28)
where Da and Da, mo are gas-phase diffusion coefficients for pesticide and water respectively; a
is a value that depends upon the conceptual model believed to describe the volatilization process
and ranges from 0.5 for the surface renewal model to 1.0 for the stagnant film model
(Cusler,1984; Schwarzenbach et al., 1993). The standard water bodies use a value of 1.0 for a
thus implying a stagnant film model; however, some laboratory data suggest that a may be better
represented by a value of 0.67 (Mackay and Yuen, 1983). The diffusion coefficient of the
pesticide is related to the diffusion coefficient of water by the common approximate relationship
(e.g., Schwarzenbach et al., 1993):
™ '° (2-29)
Substituting (30) into (29) and assuming that a is equal to one results in the following
relationship:
r is -°-5
k =k.
VH2O
MW
The resulting relationship is
ka = [0.00005+ 0.0032u01
(2-30)
(2-31)
H =
(2-32)
The Henry's Law constant is generally not available for pesticide registration, and in such
cases, it is approximated from vapor pressure and solubility as follows:
(vp/760)
(Sol/MW)
where vp = vapor pressure [torr]
sol = solubility [mg/1]
The Henry's Law constant varies with temperature according to a Van't Hoff relation as follows
(Staudinger and Roberts, 2001):
H(T) = H ,
T.
ref
(2-33)
where H(T) is the Henry's Law constant as a function of temperature
Ah = enthalpy of phase change from solution to gas [J/mol]
18
-------
R = universal gas constant =8.314 J/K/mol
Href = known Henry's Law constant at Tref [m3atm/mol]
TK = ambient (water) temperature [K]
Tref = temperature at which Href was measured [K].
The heat of enthalpy is generally not supplied for the pesticide registration process;
however, because of its important effect on volatilization and because estimation methods are
available (e.g. USEPA 2009b), it is included in PFAM. Enthalpies for pesticides are around
20,000 to 100,000 J/mol (Staudinger and Roberts, 2001; Feigenbrugel et al. 2004). The
temperature effects on volatilization dissipation are given in Figure 2-5 for several cases that
span the likely range of enthalpies for pesticides. The solvation enthalpy can have important
effects on volatilization as the figure shows. The effect of the reference Henry's coefficient and
temperature are given in Figure 2-6. Both 2-5 and 2-6 show that temperature is an important
consideration.
Aside from the temperature effects associated with the equations above, this model also
ceases volatilization if the temperature goes below 0°C, with the assumption that there will be ice
cover below 0°C which hinders volatilization. Also when depth is effectively zero (no water
volume in the water compartment) the program switches the volatilization rate to zero. If
volatilization from dry soil is an important process for a specific chemical, then model users can
incorporate the volatilization component of dissipation into the overall dry-soil degradation rate.
In this model, wind speed varies on a daily basis. The effect that wind speed has on
effective half-life is given in Figure 2-7 for a 10-cm deep pond. The figure shows that wind
speed variations will have an increasingly dramatic effect as Henry's law coefficient is reduced.
The use of daily wind speeds will thus likely have significant short-term implications (e.g., for
acute concentrations) for low Henry's law compounds.
19
-------
150
125
100,000 J/mol
50,000 J/mol
20,000 J/mol
ro
I
100
ro
o.
ro
N
•+=
JS
o
10
20 30
Temperature
40
50
Figure 2-5. Sensitivity of temperature and enthalpy of solvation (in legend) on dissipation by
volatilization. This example represents a Henry's coefficient of 10-6 atm-m3/mol and a reference
temperature of 25°C.
"ro
Q.
100
« 75
-------
100
-8
-7 -6 -5
Log Henry's Constant [log(atm m3/mol)]
-4
-3
Figure 2-7. Sensitivity of volatilization half-life to wind speed (values in legend) and Henry's Law
Constant. Simulations were created with a 10-cm pond at 25°C and a compound with a
molecular weight of 200.
2.1.3.3 Effective Benthic Region Dissipation (Ti)
The overall benthic degradation in the standard water bodies, as defined in equation (2-
8), is affected only by biodegradation and hydrolysis. As with the water column, EPA assumes
that biodegradation in the benthic region affects all forms of pesticide (both dissolved and sorbed
forms) and that hydrolysis affects only aqueous dissolved forms.
2.1.3.3.1 Benthic hydrolysis (jahydr 2)
Benthic hydrolysis is assumed to occur at the same rate as hydrolysis in the water
column, and the previous discussion of hydrolysis in the water column applies for the benthic
region. Thus,
(%dr2 = Hhydrl (2-34)
2.1.3.3.2 Benthic Metabolism (|ibio_2)
Benthic metabolism may occur under aerobic or anaerobic conditions. Either rate can be
derived from laboratory tests following standard EPA-approved protocols. These studies are
typically conducted in aqueous/sediment systems at 20 to 25°C. As with water column
metabolism, EPA assumes that sorbed-phase degradation occurs at the same rate as aqueous-
phase degradation because of the inability of the test to distinguish the two. Temperature effects
on metabolism are accounted for in an identical manner as for the water column (see previous
discussion on water column metabolism). The effective rate is thus
r^bio 2 H1 measured
lief
Q
10
(2-35)
where ^measured = laboratory measured anaerobic metabolism rate at Tref
21
-------
T = temperature of modeled water body [°C]
Tref = temperature at which anaerobic laboratory study was conducted [°C].
2.1.3.3.3 Dry soil degradation
When water level is effectively zero, the model provides for a separate input to account
for unflooded soil degradation. Typically this will be taken from an aerobic soil degradation
study following standard EPA-approved protocols. Under unflooded conditions equation 2-35
still applies, but the measured value (^measured) will be taken from the aerobic soil studies.
2.1.3.3.4 Benthic Leakage Coefficient (A)
The leakage coefficient in the benthic region represents the flow through the benthic
region. Unlike in the water column, the benthic region concentration can increase or decrease
due to leakage, depending on the relative concentrations in the water column and benthic
regions. It has a similar effect on the benthic region as the mass transfer coefficient does, as
evidenced by its position in equation 2-6. This parameter can be readily calculated from
equation 2-9 or equivalently, as would be done in a computer program, by using previously
calculated terms as in the following:
(2.36)
2.1.3.4 Mass Transfer (Q)
The mass transfer term is best thought of as an overall coefficient that includes all means
of pesticide exchange between the water column and benthic regions. This includes exchange
through the aqueous phase as well as by mixing of sediments between the two compartments.
The physical process of this combined mixing is assumed to be completely described by a first-
order mass transfer coefficient (co). The parameter co is referenced to the aqueous phase, but
implicitly includes exchange due to mixing of sediments as well as aqueous exchange. In
compartment modeling, it is unnecessary to explicitly model the individual exchange
mechanisms since all phases of a pesticide within a compartment are at equilibrium and therefore
the concentration of pesticide in any given form (aqueous or sorbed) dictates the concentration of
the other forms of the pesticide.
As developed elsewhere (USEPA, 2004), the volumetric mass transfer equation (2-9) can
be broken into somewhat more fundamental terms as follows:
•fKedA^ + vJ (2-37)
d2
And therefore
Q = -^- (2-
d2
38)
where d2 is the benthic depth, and where the term kxfer is a geometry-independent mass transfer
coefficient [m/s]. This latter term is best viewed as an empirical estimator of overall water
column to benthic mass transfer. The term kxfer is on the order of 10"8 m/s according to several
sources (Vanderborght and Wollast, 1977, Schwarzenbach et al., 1993, Burns, 2000).
22
-------
2.1.4 Pesticide Applications
PFAM allows the user to apply pesticides in ways that should cover most application
possibilities. Users may apply pesticide to the dry soil or to the flood water. Additionally users
may specify that the pesticide is manufactured to be slowly released into the application area.
Dry applications will occur if the user specifies that the pesticide is applied to an unflooded field.
In this case, the pesticide is automatically applied to the soil, which becomes the benthic region
upon flooding. Upon flooding, the pesticide may enter the water column through physical mass
transfer processes. If the user applies pesticide during a flood, then all pesticide is initially
placed into the water column. This latter application occurs regardless of the presence of a
canopy. Canopy interception does not occur in PFAM because the required foliar degradation
and washoff parameters are typically unavailable for pesticide assessments. Thus, until better
foliar studies and better data become available, PFAM makes the environmentally protective
assumption that all pesticide enters the water column when an above canopy application occurs.
When the slow release option is selected the pesticide is assumed to be released in a first-
order manner in which the amount of pesticide unreleased is
Mu =M0e-ksrt (2-39)
where Mu is the mass of unreleased pesticide (kg), Mo is the original amount of pesticide (kg), ksr
is the release rate (day"1) and t is time (days). PFAM calculates the mass released each day by
(2-40)
where Mt is the mass release for time t and At is the time interval (1 day). For practical purposes,
the mass released does not extend to infinity. Rather PFAM allows the slow release to occur
until 95% of the pesticide is released and the remaining (5%) is applied on the following day.
When multiple years are simulated, PFAM will automatically apply the pesticide in the same
manner for all years. This practice is in keeping with the standard way that the US EPA
performs exposure assessments for pesticides.
2.1.5 Degradates
Degradates are handled exactly like the parent in regard to their transformations. The
production of degradates is from the first-order degradation of the parent compound and can be
due to water, dry soil, or benthic metabolic degradation, photolysis or hydrolysis. Users can
specify the stoichiometry of the degradate production. Up to two degradates in series are
possible with PFAM as in
P -> XDl-> Y D2
Where P is the parent compound, Dl is the first degradate, X is the number of moles of Dl
created when one mole of P degrades, D2 is the second degradate that forms by the degradation
of Dl and Y is the number of moles of D2 formed for one mole of Dl degraded. The molar
ratios should be available from the stoichiometric equations supplied by the pesticide study
submissions.
23
-------
2.2 Computations
Because of the advantages of using an analytical solution for the chemical concentrations,
the model is solved in a daily piecewise fashion. This is achieved by approximating the water
volume changes by discrete daily changes in which the volume of the water column changes at
the beginning of the day and remains constant for the duration of that day, as shown previously
by equation 2-1. With the approximation of constant within-day volume, the concentration
calculations are amenable to an analytical solution for the daily time steps. Mass is conserved in
the water column by recalculating a new beginning day concentration with consideration of the
volume change.
2.2.1 Initial Conditions
Initial concentrations for the standard water bodies are determined by the pesticide mass
inputs. Depending on the pesticide-management practice, a pesticide may be applied during a
flooded condition or directly to the ground prior to flooding. For pesticide applications during a
flooded period, the model places all applied pesticide into the water column. For pesticide
applications during dry periods, the model places all pesticide mass into the soil compartment.
For this model, there is an instantaneous water depth change at the beginning of the day
due to hydrologic conditions (see Flood and Overflow Control above), and the pesticide
concentration in the water column is adjusted accordingly. The initial concentrations, upon
addition of new pesticide inputs, are then expressed as:
prior s-i
10, prior
f
J wl, prior
(2-41)
(2-42)
where Minput,i = mass of pesticide applied to water column (kg)
MmPut,2 = mass of pesticide applied to benthic/soil compartment (kg)
Cio = initial aqueous concentration of water column for current time (kg/m3)
€20 = initial aqueous concentration in benthic region for current time (kg/m3)
Cio,Prior = aqueous concentration in water column before new mass additions (kg/m3)
C20,pdor = aqueous concentration in benthic region before new mass additions (kg/m3)
vi, prior = the water column volume from the previous day (m3)
fwi!prior = fwi from the previous day
2.2.2 Analytical Solution for Concentrations
Equations (2-5) and (2-6) along with the initial conditions represent the two equations
describing the standard water bodies. These equations are in the form of
(2-43)
(2-44)
dt
where:
24
-------
= Q0
F=-n-n-A
These equations have the following solution:
l
' B
B
where:
F)2-4(FA-BE)
F)2-4(FA-BE)
(2-45)
(2-46)
(2-47)
(2-48)
(2-49)
(2-50)
(2-51)
(2-52)
(2-53)
(2-54)
Average concentrations can be determined over any interval in which all parameters
remain constant. In the case of the proposed model, parameters change on a daily basis, so the
average water column concentration over any of these time intervals, is expressed as
(2-55)
where Ci,avg = average water column concentration from ti to \.i [kg/m3]
ti = beginning of the time interval [s"1], (zero for the case of daily estimates)
t2= end of the time interval [s"1], (86400 seconds for PFAM case of daily estimates)
2.3 Post Processing
Effluent from PFAM can be optionally routed to various user-defined water bodies,
including two standard EPA water bodies: the EPA Index reservoir and the EPA farm pond.
Both flowing and static water bodies may be simulated. The receiving water body hydrology
and chemical processes are calculated by a program nearly identical to the PFAM program
already described. The receiving water body, however, receives influent water from PFAM
effluent as well as runoff from the surrounding area.
Receiving water body possibilities are shown in Figure 2-8. In this figure, the receiving
bodies (or mixing cells) are depicted by the blue cubes. These mixing cells, with user-defined
dimensions, represent the outlet or terminal point of a watershed and could represent a section of
a flowing water body or a large reservoir, depending upon the parameterization. The
25
-------
postprocessor automatically calculates the runoff from the water shed and tracks the effluent
from the flooded plots within the watershed. Influent water into any receiving water body is
determined by the following equation:
Where
(2-56)
Q = flow entering the receiving water body
QPFAM = PFAM water releases
Qws = Runoff from surrounding watershed
B = Baseflow through water body
P = Direct Precipitation-Evaporation
The PFAM releases are calculated as previously described. PFAM will give the daily amount of
water that leaves the system, and this release feeds into the receiving water on a daily basis. The
runoff from the surrounding area is calculated by the NRCS curve number method (NRCS,
2003). This calculation requires a user estimate of the watershed area and an appropriate curve
number. A base flow may also be appropriate and the post processor allows for this entry. This
base flow value should represent the flow through the receiving water body during periods not
dramatically affected by local rain event or flooded agriculture releases.
Chemical processes are calculated the same way as described in PFAM above with the
same chemical inputs. Concentrations in the receiving water body are reported in a manner
similar to the way that the US EPA reports concentrations for standard pesticide assessments
(i.e., as l-in-10-year events for the peak, 4-day, 21-day 60-day and yearly averages).
Flowing Water Body Section
Watershed Boundary
Figure 2-8. The conceptualization of the water body and watershed. Various setups such as these are
available by use of the post processor.
26
-------
3. Evaluation of PFAM Using Criteria for Regulatory
Assessments
In general, pesticide-exposure models should be able to reasonably represent pesticide
behavior, capturing the most environmentally salient physical and chemical properties of
pesticide use. At the same time, such models should not be overly complex because only a few
chemical properties are available from the pesticide registration process. Furthermore, pesticide
exposure assessments are typically generic with regard to their representation of the
environment. In other words, the model scenarios are often surrogates for large areas or an entire
nation and are rarely site-specific. On a large spatial scale with commensurate large-scale
variability, it may not be productive to create complex models and populate them with site-
specific parameters since such efforts would not raise the accuracy of the model above the
background noise. However, because regulatory pesticide exposure models are primarily used
for environmental protection, they should provide reasonable high-end estimates of
environmental concentrations when appropriately parameterized and compared to a sampling of
field data. The purpose of this investigation is to evaluate PFAM in the context of a regulatory
application by comparing the model-predicted results with actual field data and evaluate whether
PFAM can produce high-end, but reasonable estimates.
PFAM was developed consistent with the guidance documents of the USEPA Council for
Regulatory Environmental Models (CREM) (USEPA, 2009a) and the USEPA quality assurance
(QA) program (USEPA 2002). The CREM guidance (USEPA, 2009a) covers development,
evaluation, and use of models intended for environmental regulatory decisions. The QA
guidance describes specific information concerning what is required to plan for model
development to ensure that a model is scientifically sound, robust, and defensible for regulatory
purposes. These two guidance documents are complementary and provide a solid basis for
model development and application in a regulatory setting.
According to USEPA (2009a), model evaluation is the process for determining whether a
model's results are of sufficient quality to serve as the basis for a decision, where the meaning of
quality depends on the model's application or intended purpose and is defined by quality
objectives. Evaluation addresses whether model development incorporates sound science,
whether the model requirements are suitable for the available data, and whether the model
compares sufficiently well with real data. Model verification (code checking) has been
performed throughout PFAM development and has been documented by US EPA Science
Advisory Panels (USEPA, 2004) and by other agencies (Luo, 2011). The model's quality
objectives concerning data corroboration, often referred to as validation (Rykiel, 1996), are
covered in this chapter.
The quality objectives of regulatory models such as PFAM, which are surrogate
representations of pesticide use over large spatial scales, focus on their performance as
screening-level tools. Because of the high uncertainty associated with large-scale assessments, a
regulatory model's output should provide reasonable but appropriately conservative estimates of
exposure in order to be protective. That is, the quality objectives are (1) that the possibility of
incorrectly giving passage to a chemical that is dangerous is at an acceptable level and (2) that
27
-------
there is not an unreasonable burden on the regulated community due to excessive over
predictions of exposure.
How well a screening model minimizes the potential for false negatives (the first
objective) can be evaluated by comparing model estimates to measured concentrations in the
field. Determining whether model estimates do not place undue burden by excessive false
positives (the second objective) is less straightforward. One evaluation measure for the second
objective is to compare PFAM's predicted concentrations to that of the currently used and
presumably more conservative screening model. In this sense, the second objective is met if the
model predicts concentrations that are less than the currently used screening model, thereby
allowing passage of chemicals that otherwise would require more testing
3.1 Methods
3.1.1 Field Studies
Field data used in this evaluation came from pesticide registration studies submitted to
the United States Environmental Protection Agency (USEPA). Manufacturers interested in using
their pesticides for aquatic agriculture or other aquatic applications may submit aquatic field
studies that characterize the fate of a pesticide after its application to a water body such as a rice
paddy or pond. Typically, these studies follow the course of a pesticide's existence at the study
site, from the time of its application until either the pesticide completely disappears from the site
or the water body completely drains.
These studies vary widely in quality and usability for PFAM evaluations. Out of the
hundreds of aquatic field studies, five were selected based on the completeness of site
characterization, the temporal and spatial resolution of the pesticide measurements, and the
persistence of the pesticide. Complete site characterization requirements include daily weather,
soil and sediment characterization, and water levels. The need for temporal and spatial
completeness of the data required that both sediment and overlying pesticide concentrations were
measured frequently enough over a long enough period such that important hydrological and
chemical events would be captured (i.e., flooding, draining, as well as pesticide degradation).
With regard to pesticide degradation, somewhat persistent pesticides typically gave better
temporal resolution than fast-degrading pesticides. Studies with pesticides that disappeared
within a few days lacked enough temporal data points to be useful. In addition, the final chosen
studies exhibited a range of aquatic management practices, in order to allow examination of a
variety of pesticide-application schemes, including preflood applications, post-flood
applications, flow-through systems, and static ponds. Table 3-1 provides a summary of the
relevant characteristics of the five selected studies.
These five studies (Table 3-1) represent a range of potential applications. Four of the
studies are for rice applications, which will likely be the most frequent application for PFAM,
and one study is for aquatic weed control in a small pond. Study A represents a rice culture
application in which the pesticide is applied to the soil before the field is flooded, as is often the
case for weed control. Studies B and D represent cases in which a pesticide is applied to a
flooded field, which is common for fungicides or insecticides. Study C represents another
common rice application involving a continuous flow system in which flowing water is
maintained at a constant level by a weir. In addition to these rice applications, Study E provides
an example of a natural pond-type application of pesticide. The next sections further describe
the details of the studies.
28
-------
Table 3-1. Field Properties relevant to simulation.
Property
Location
Water Management
Practice
Chemical
Area (m2)
Organic Carbon (%)
Application date
Pesticide Mass Applied
(kg/ha)
Flood level (cm)
Tolerance (cm)
Flood Date
Drain Date
Turn Over (d'1)
Crop Plant Date
Crop Full Height Date
Crop Harvest Date
Max Crop Coverage
(fraction)
pH
SS (mg/L)
Bulk Density (g/ml)
Study A <»>
Stutgart, AR
Preflood
Application
Pendimethalin
2800
1.1
June 2
1.1
7.6
5.1
June 22
Sep20
-
May 29
August 8
not harvested
0.90
8
-
~
Study B
Bay City, TX
Continuous
Flow
Propiconazole
7100
1.55
Sep 5 & 20
0.189
7.6
-
-
-
1.06
Augl5
Octl
Oct30
0.90
7.2
-
1.35
Study D
Elk Grove,
CA
Static Pond
Triclopyr
1200
0.88
July 26
20
80
80
~
~
~
~
~
~
0
8
8
1.1
(a) US EPA (1998a), (b) US EPA (1999), (c) US EPA (1992), (d) US EPA (1994), (e) Petty et al. (2001)
3.1.1.1 Study A: Stuttgart AR Pendimethalin Study Summary
This study (USEPA, 1998a) is an example of a pre-flood application. The study took
place near Stuttgart, Arkansas on two dry-seeded rice plots. The soil was a silt loam with
organic carbon content of 1.1 percent. Each plot was about 1400 m2. The study reflected typical
rice agronomic management practices for the region, including flooding of the field, adding
makeup water, and eventual draining of the floodwater. Table 3-1 presents the characteristics
that are relevant to modeling.
Because the two plots in this study were nearly identical, samples from the two plots are
treated as replicates rather than as different studies. The two plots were separated by about 30
feet. The plots differed in the time that the rice crop was planted (May 11 and May 29), but were
otherwise equivalent with regard to management practices. Note that planting date (and
effectively canopy development) only affects the photolysis rate in PFAM. Because degradation
29
-------
by aerobic metabolism (as discussed later under Chemical Parameters) will overwhelm
photolysis in this case, the model will be insensitive to planting, thus allowing the plots to be
treated as replicates.
Plots were flooded on June 22 and maintained in flood stage until September 20. The
flood level varied from 5 cm to 10 cm (average 7.6 cm). Water was supplied as needed,
typically in amounts of about 5 cm at a time, with the total estimated water input ranging from
60 cm to 84 cm for the duration of the study (ending Sep 9). Pendimethalin was applied as a
broadcast spray by a backpack sprayer at 1.1 kg/ha to dry plots on June 2. Rice plants were at
early germination stage to 4-leaf stage for the two plots, so plant interception of pendimethalin
was minimal.
Precipitation, temperature, and wind speed were measured at an offsite weather station
about eight miles from the site. Evaporation was not included in the data, so evaporation was
estimated by Ham on's formula (Ham on, 1961) which bases evaporation on temperature and
latitude. The study report recorded the weather during this period to be typical.
Samples were taken from both the water column and the soil (before and after flooding).
Soil/sediment samples were taken from 0 to 15 cm deep and divided into subsamples of 0 to 7.6
cm and 7.6 to 15 cm, with pesticide mass measured for each subsample. The vast majority of
benthic pesticide mass (96 to 100%) remained in the top 7.6-cm meter core. Water samples were
taken as grab samples after flooding. Sampling continued until near complete dissipation of the
chemical.
3.1.1.2 Study B: Carlisle AR, Bispyribac Sodium Summary
This study (US EPA, 1999) is an example of a post-flood application of pesticide; that is,
the farmer first flooded the field and then applied pesticide directly to the water. The study took
place on a 94-m2 plot in a rice paddy near Carlisle, Arkansas. The soil was a poorly drained silt
loam with an organic carbon content of 0.57 percent. Table 3-1 summarizes relevant modeling
parameters. For this Arkansas study, typical agricultural practices were used, including flooding
and draining of the field.
The field was flooded May 31 to a depth of about 15 cm. Bispyribac sodium was applied
to the flooded field on June 12 at 0.06 kg/ha to the plot by a backpack sprayer. At application,
the rice was about 0.42 m high with a 50% canopy cover. For the duration of the study,
irrigation water was added as needed in about 7.6-cm amounts (tolerance was thus assumed to be
7.6 cm).
Water samples were taken up to 56 days after application. Sediment samples were taken
up to 112 days after application from 0-15 cm and from 15 cm to 30 cm. No pesticide was
detected below 15 cm. Because soil samples were drained of excess water, there is potentially
some added uncertainty regarding mass of pesticide lost. However, the samples remained quite
wet after draining (25% by weight) so that losses would be fairly negligible. Sampling continued
until depletion of the chemical.
Weather, rain, and air temperature were recorded on site. Pan evaporation was taken
from the US EPA data set (Burns et al., 2007) for the nearby Little Rock weather station. The
study kept records of the recharge water used to maintain field flood level, but the study did not
directly track water level in the field. The total recharge water was 310 cm through day 112.
30
-------
3.1.1.3 Study C: Bay City, TX Site Propiconazole
This test (US EPA, 1992) is an example of a continuous flow-through system. In this
type of system, the water level remains fixed by water continually flowing through the system
and over a weir, a practice that is typical for this Texas area. The study took place on a 0.71-ha
plot near Bay City, Texas. The average turnover due to flow through the water body was 1.06
volumes per day. The depth of the water was 15 cm. The sediment was clay with an organic
carbon content of 1.55%.
The plot received aerial applications of propiconazole at 0.189 kg/ha on September 5 and
again on September 20 to the flooded field under calm winds (0-3 mph). The original rice crop
had been harvested on August 1, and the pesticide applications were applied during the ratoon
crop. The applications occurred about mid-way through the ratoon-growing season, allowing for
the possibility of canopy interception of propiconazole; however, the study did not report the
amount of canopy coverage. Because PFAM does not account for canopy interception, this
study will provide for some information concerning the interception significance in regard to
PFAM output. Above-canopy application efficiency was reported to be 59% and 75% for the
two applications, respectively. The cause of these inefficiencies is unknown. The field was
drained over the period from Sep 30 to Oct 2.
Weather measurements were taken on site and included temperature, precipitation, and
wind speed. Evaporation was not recorded on site, so it was taken from US EPA data (Burns et
al., 2007) for the nearest available site which was Victoria, TX. A digital flow meter kept track
of outflow from the water body.
Both sediment and water samples were collected in the paddy. Soil/sediment samples of
at least 25 cm were taken from three locations on each sampling day. Sediment samples were
taken and composited from depths of 0 to 10 cm and from 10 to 20 cm. No pesticide was
detected below 10 cm. Water samples of 1.3 liters were taken at three locations on each
sampling day and composited into a single sample. Water quality measurements such as pH and
dissolved oxygen were also recorded.
3.1.1.4 Study D: Pattison Texas, Carbaryl
This study (US EPA, 1994) is another example of a post-flood application of a pesticide
in which the pesticide is applied directly to the water of a flooded field. This study took place
near Pattison, TX. The water level varied over the course of the study but was typically about 7
cm deep. The plot area was 929 m2, with levees surrounding it to retain water. The soil in the
plot was a silt loam with an organic carbon content of 0.66%.
Rice planting occurred on April 15, followed by flooding of the plot on May 28.
Carbaryl was applied by spray boom above the rice canopy two times, once on June 18 and again
on June 23. At the time of the applications, canopy coverage was about 50 percent, so canopy
interception of carbaryl should have occurred. However, as stated in Young (2012), foliar
washoff and foliar degradation data is almost never available for pesticide assessments. For this
reason, PFAM developers elected to take a conservative approach and allowed only direct
application of pesticide to the water body rather than to the plant canopy (Young, 2012). This
study will allow a test of the protectiveness of that assumption.
The site was equipped with a siphon mechanism that maintained a minimum water level
in the plot. The water level varied over the course of the study with a maximum of 12 cm. The
minimum allowed depth was around 2.5 cm. It was clear from the water level data that the water
level was allowed to rise during the study from 4 cm to 12 cm. The minimum acceptable level
31
-------
also appeared to rise, although these management practices were not detailed in the study report.
Because the exact water level was unknown and because of its potential impact on the results,
the water level for the simulation was calibrated to roughly approximate these changes in depth
by changing the simulated depth and tolerance three times during the simulation. The plot was
drained on July 30, and the rice was left standing in the plot for the duration of the study.
Both soil and water samples were taken during the course of the study and analyzed for
carbaryl. Soil was sampled to 15 cm and segmented into two 7.5-cm segments. Top segments
were combined to create a single sample. Carbaryl did not move below the top 7.5-cm segment.
The study reported weather data, including daily rainfall, pan evaporation, and air temperature.
The study did not report wind speeds, so wind speed was obtained from the closest NOAA
weather station (Houston-Bush Airport).
For carbaryl, system pH is very important for degradation by hydrolysis. Carbaryl
degradation is highly dependent on pH, with the degradation rate increasing as pH increases. For
the measured water pH of 7.9 for this system, the half-life of carbaryl was estimated to be about
3 days based on data in US EPA (2007b).
3.1.1.5 Study E: Sacramento, CA Triclopyr
This study (Petty et al., 2001, USEPA 1997a) was conducted near Elk Grove, CA at the
California Department of Fish and Game Aquatic Toxicology Lab in two fabricated ponds. The
test ponds were 0.12 hectare with a depth of 0.8 m. The sediments were generally sandy clay
loams with an average organic carbon content of 0.88%. This study is different from the other
studies in that it is a non-crop water body without water release or resupply. This study uses the
chemical triclopyr, which is highly susceptible to photodegradation; hence this study will allow
for the evaluation of the photolysis routine in PFAM. The photolysis routine depends upon the
actual latitude of the study (38.25 N latitude) as well as factors responsible for light attenuations,
such as suspended solids. Table 3-1 reports the properties obtained from the study and used as
model inputs.
Using a 20-liter powered sprayer with a hand wand, triclopyr was applied to the ponds to
achieve a concentration of 2.5 mg/L in the water. The pesticide was sprayed slightly above or
just within the water surface. Application to both test ponds occurred the morning of July 26.
A weather station at the site collected weather data, including air temperature, relative
humidity and wind speed. Daily evaporation was estimated from temperature using Hamon's
formula (Hamon, 1961). No precipitation occurred during the study period, and during the
course of the study the ponds did not receive additional water to offset evaporation.
Water samples were taken in duplicate at two depths at each water sampling station, and
a 400-mL grab sample was collected at each sampling event at 1/3 and 2/3 total depth of the
water column. Sediment samples of approximately 300 g were collected from the top 5 cm of
benthic sediment, using clamshell post-hole diggers.
3.1.2 Chemical Properties
Chemical properties required to populate the model (i.e., sorption, degradation) are
readily available from studies submitted by pesticide manufacturers to support pesticide
registrations. These laboratory studies are conducted on soils and in environments that are not
necessarily representative of the field conditions, and thus the laboratory-derived chemical
properties values may vary from the actual site values. Direct chemical property measurements
from a particular field study site are usually nonexistent, as measurement of those properties is
32
-------
not a requirement for field-study submissions. Even though direct measurements are not
available, simulations can still be made using the best available resources. This approach would
be similar to the way a regulator would make a pesticide assessment as well, since only
laboratory-derived, non-site specific chemical properties are available for registration. This
limitation is not particularly disadvantageous, since regulators are not typically concerned with
the concentrations at any particular single site, but instead are concerned with the broader
question of concentrations for all potential use sites.
Whenever possible, chemical inputs were taken from easily accessible publically
available sources such as internet-accessible databases of pesticide registration documents or
other public databases. Occasionally when chemical parameters were not available through
publically available documents, alternate means of estimations such as EPI Suite (USEPA
2009b) were used. No effort was made to make the input parameters conservative (protective or
worst-case) as would be typically done in a purely regulatory use of an environmental model,
since application of conservativeness into the assessment is a policy decision for which none has
been made at this time. Instead, since the purpose here is to evaluate the model's performance
rather than produce regulatory values, best estimates (which for the most part, were assumed to
be the mean value of any values found) were used. Table 3-2 presents these generic non-site-
specific properties used for the PFAM simulations.
Table 3-2. Relevant chemical properties of compounds in the field studies.
Property
Study
PC Code
Molecular Wt
Soil Aerobic Half -life
(day) /(°C)
Water Aerobic Half-
life (day)/(°C)
Water Anaerobic Half-
life (day)/(°C)
Photolysis Half -life
(day)/( °N latitude)
Hydrolysis Half-life
(day)
Vapor Pressure (torr)
Solubility (mg/L)
Heat of Henry (J/mol)
Kd (ml/g)
Koo (ml/g)
Pendimethalin3
Study A
108501
281
126/25
10/25
50/20
21/42
stable
3e-5
0.375
62,000b
~
15000
Bispyribac
Sodium0
Study B
078906
452
19/25
64/25
99/25
stable
stable
le-7
73000
~
0.6-2.0
114
Propiconazoled
Study C
122101
342.2
53/25
426/25
363/25
stable
stable
le-6
110
45,000b
~
648
CarbaryP
Study D
056801
201
4/20
stable
72/20
21/40
1.67 @
pH7.9
1.36e-6
32
58,000b
~
198
Trichlopyr
Acidf
Study E
116001
256
13/25
142/25
1300/25
0.6/42
stable
1.26e-6
435
~
0.6
~
(a)US EPA (1997b), (b)US EPA (2009b),
(1998b)
(c)US EPA (2001), (d)US EPA (2006), (e)US EPA (2007b), (f)US EPA
3.1.3 Simulations and Comparisons
Most PFAM inputs can be readily determined from the field study reports or from the
methods for obtaining chemical properties as described above and given in Table 3-1 and 3-2.
Some parameters were not available from the studies, and in those cases, the PFAM default
33
-------
values were used. These defaults are the same as those used in EPA's standard water bodies
(USEPA, 2004) and are listed in Table 3-3.
Table 3-3. Default Parameter Values for PFAM (parameters defined in Young 2012)
Parameter Value
Mass Transfer Coefficient (m/s) 10~8
Benthic Depth (m) 0.05
Benthic Porosity (m3/m3) 0.5
Bulk Density (g/cm3) 1.35
foe Benthic (—) 0.01
foe Water Column (—) 0.01
Suspended Solids (mg/L) 30
Chlorophyll (mg/L) 0.005
Water Column DOC (mg/L) 5
Benthic DOC (mg/L) 5
Q10 2
DFAC 1.19
Water concentrations and water levels were simulated with PFAM and compared to the
available data. Comparison of water-column concentration was straightforward by observing the
PFAM output of daily average concentrations along with the available study data. As for water
management, PFAM simulates water additions, releases, and overflows, and level; however, the
studies did not always report this information, so only limited direct comparisons could be made
regarding hydrology (typically only water additions were reported). Comparisons were made
when these data were available.
Soil concentration comparisons are presented here in terms of mass per area since the
field sampling strategy did not indicate how the pesticide was distributed within the sample core.
Thus, it is not possible to determine a volumetric concentration that is comparable to PFAM
estimates that are based on a 5-cm sediment zone (see Table 3-3). This strategy of comparing
the mass-per area values of the PFAM simulation with the sample core data allows an evaluation
of how well PFAM simulates the total mass of pesticide in the sediment rather than the
concentration.
3.1.4 Hypothetical Regulatory Simulations
Historically, the US EPA bases regular (non-flooded agriculture) aquatic assessments of
pesticides on a 30-year simulation in which a pesticide is used at its maximum application every
year. This type of assessment allows for the analysis of the temporal variability, which is
primarily due to changes in the weather from year to year. In order to evaluate the potential use
of PFAM in such a likely regulatory application, PFAM simulations were also conducted with a
30-year simulation. These simulations were conducted using the scenarios and chemical
properties previously described but with the USEPA regional rainfall data (Burns et al., 2007)
and with the additional condition that the same application pattern was made every year of the
30-year simulation. Table 3-4 presents the identification information for the weather data (Burns
et al., 2007) used in the long-term simulations.
34
-------
Table 3-4. Weather data used for long-term simulations.
Study
A
B
C
D
E
Actual Location
Stuttgart AR
Carlisle, AR
Bay City, TX
Parti son, TX
Sacramento, CA
CEAM Weather File
Little Rock (1961-1990)
Little Rock (1961-1990)
Victoria (196 1-1 990)
Houston (196 1-1 990)
Sacramento (1961-1990)
WBAN
13963
13963
12912
12960
23232
Chemical and scenario information were kept the same as in the field-study-comparison sections
described above. Because there is yet no guidance or policy regarding PFAM inputs, no effort
was made to make the chemical input parameters conservative, as is done in other US EPA
aquatic assessments (e.g., using 90th percentile degradation half-lives as inputs rather than typical
values). Thus, there is no built-in conservatism for the evaluations that follow.
In addition to the long-term analysis of PFAM variability, comparisons with the current
first tier screening calculation was also conducted. This first-tier screening concentration is
equivalent to the concentration that would occur if the pesticide application were equilibrated
with 10 cm of water and 1 cm of soil that has a bulk density of 1.3 g/mL and 1% organic carbon.
The resulting Tier 1 concentration estimate is only dependent on the partition coefficient and is
calculated as follows:
w "0.00105+0.00013^
Where Cw = the Tier 1 water concentration
Ma = mass of pesticide applied per area (kg/ha)
Kd= distribution coefficient (ml/g) = 0.01KOC
Koc = organic carbon partition coefficient (ml/g)
3.2 Results and Discussion
3.2.1 Study A: Arkansas, Pendimethalin Simulation Results
Figure 3-1 shows the PFAM-simulated water levels for the Study A site along with the
reported amounts of water additions. This study reported water additions but not water levels.
Therefore, hydrology comparisons focus on the additions rather than on water level. As Figure
3-1 shows, the predicted water additions match well in frequency and magnitude to the actual
water additions, except for the period between 40 and 50 days. Between 40 and 50 days, there
were two more actual additions than what PFAM predicted. Primarily because of the refills
during this 10-day period, which amounted to about 1 m, the total PFAM estimates (0.52 m) are
lower than the actual total water additions 0.71 m (±0.12). The discrepancy is possibly due to
the offsite measurements of the weather data, which may not exactly correspond to the local
weather at the site.
Figure 3-2 shows the predicted water column concentration for Study A along with the
measured data. Note that this was a preflood application of pesticide, and therefore water
column concentrations do not appear until after the field is flooded (20 days after pesticide
application). The actual water column concentrations do not achieve the initial PFAM-simulated
value. The most likely reason for this is that the simulation is also over predicting the available
35
-------
pesticide in the soil at the time of the flood, which may be due to an underestimation of the soil
degradation rate, an overestimate of the actual amount of pesticide applied, or possibly extraction
and recovery issues. Note that chemical fate parameters did not come from direct field
measurements at this site, as described in the Chemical Properties section. As a result, there is
uncertainty regarding the accuracy of the soil degradation estimates. Nevertheless, the
predictions are on the same order of magnitude as the data, and as is desirable for a regulatory
model, the simulations err on the high side of the data. One additional noticeable quality of the
PFAM prediction is that it fluctuates considerably and coincides with the fluctuation in the water
level. In this case, water level decreases cause increases in concentration, as water is volatilizing
at a faster rate than the pesticide.
Figure 3-3 shows the simulated benthic concentrations. The initial measured mass is
similar to the simulated mass, as would be expected from an acceptable field study. (This result
indicates good recovery of pesticide from the field). PFAM predicts the remainder of the data
reasonably well, although PFAM concentrations are consistently higher than the data, which is
an acceptable quality for a regulatory model. Actual dissipation occurs somewhat faster than
PFAM predictions, and this outcome may be due to the underestimation of degradation rates.
Again, as with all the studies here, the environmental fate properties did not come from direct
measurements at this site and thus the actual degradation rate at this site is unknown.
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
- Simulated Water Level
Actual Water Additions
• Simulated Water Additions
\
I
20
40 60 80
Days after Application
100
120
Figure 3-1. Simulated water level of Study A along with the reported and simulated water
additions.
36
-------
20 40 60 80 100
Days after Application
120 140
Figure 3-2. The simulated floodwater pesticide concentration and the measured data for Study
A. (Water column concentrations do not exist until after flooding on Day 20.)
160
140
CM
j= 120
100
o
O
80
60
40
20
i a
a.
0 20 40 60 80 100 120 140
Days after Application
Figure 3-3. The simulated sediment pesticide concentration and the measured data for
Study A.
37
-------
3.2.2 Study B: Arkansas, Bispyribac Sodium Simulation Results
Figure 3-4 shows the simulated water levels and the reported and simulated water
additions. Like Study A, this study gave water additions but not water levels over time.
Comparisons thus focus on these additions rather than the direct water levels. The actual water
additions are very similar in magnitude and timing to the simulated additions, inferring that
PFAM captures the hydrologic functions reasonably well. Predicted total additions were 0.58 m,
which was about 20 percent higher than the actual reported additions of 0.48 m. This is a
reasonably good simulation especially since pan evaporation was not measured on site but
instead taken from a station 40 km away.
Figure 3-5 shows the water column concentration over time. Note that this was a
preflood application, so water column concentrations do not appear until the flood occurs on day
11. The simulated initial concentration (day 11) is similar to the measured concentration,
indicating good application efficiency. This good initial simulation occurs despite the presence
of a canopy with 50% coverage, which supports neglecting pesticide canopy holdup, as does
PFAM.
With regard to the pesticide's temporal decline in the water column, the simulated
concentration drops slower than the measured data, indicating that the actual pesticide moves
into the soil faster or degrades faster than the simulation. Note also that the simulation clearly
pulsates due to water level changes; however, the data have lower temporal resolution and thus
do not confirm this effect.
Figure 3-6 shows the simulated and measured benthic concentration over time. Both the
simulated concentration and the data exhibit an increase, a peak, and a decline over time. The
increasing benthic mass is initially well simulated, but the actual data peak and begin to decline
much sooner than the simulation. The actual degradation rate in the sediment appears to be
faster than the simulated rate, which is not unexpected given that the degradation rates were not
measured at this particular site. However, PFAM does err on the high side, which is an
acceptable feature in a regulatory model. The apparent poor representation of the benthic mass
should be placed into context: the amount of pesticide mass in the benthic region at any time
according to the data is about 0.1% of the total mass in the system. Because of the relatively
small amount of mass transferred to the benthic region, it is reasonable that there should be a
large uncertainty in the measurements.
38
-------
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
\
\
\
\
s
\
\
\
Ds,
\
\
\
\
\
\
\
Simulated Water
V
\
\
\
PA
\ .
A
\
Level
tions
Additions
l\
\
\
\
0 10 20 30 40 50 60 70 80 90
Days After Flood
Figure 3-4. Simulated water level of Study B along with the reported and simulated water
additions.
70
60
o) 50
w 40
8
B
I
10
0
— Simulation
O Data
o
8
-e-
0 20 40 60 80
Days After Flood (12 days prior to Pesticide Application)
Figure 3-5. Simulated flood water concentration and measured data for Study B.
39
-------
O)
£=
O
0)
O
c
O
O
T3
0)
1.2
1 -
0.8
0.6
0.4
0.2
0
0
— Simulation
O Data
O
20 40 60 80 100
Days After Flood (12 days prior to Pesticide Application)
Figure 3-6. Simulated sediment concentration and measured data for Study B
120
3.2.3 Study C: Texas, Propiconazole Results
Study C site was a flow-through system in which a weir held water levels constant, so
hydrologic simulations are not provided here. The constant flow occurred at a rate of 1.06
volumes per day, resulting in a water column half-life of about 0.65 days. Because of the very
long half-lives due to degradation of the applied pesticide (hundreds of days, Table 3-2), washout
is expected to be the dominant source of dissipation in the system. Figure 3-7 shows the fast
washout-driven dissipation.
Figure 3-7 also shows the water column concentrations for Study C. The simulated
initial concentration is about five times higher for the first application and about twice as high for
the second application. These discrepancies are due in part to the low reported pesticide
application efficiencies for the study (59% and 75%, respectively). The differences could also be
due in part to canopy interception, but this possibility cannot be verified. (The study reported that
a rice crop was in place, but it did not report the canopy coverage). The PFAM simulation
captures the long-term concentration better than the short-term concentrations although the
measured concentrations appear to drop off slower than PFAM predicts. This additional
discrepancy supports to some extent the possibility of pesticide hold up and subsequent slow
washoff from the foliage into the water column. Canopy holdup and washoff are areas needing
additional research for models such as PFAM, but including those processes without
commensurate data would be guesswork and inappropriate in an environmental protection
context. Nevertheless, PFAM does give protective concentrations for the bulk of the simulation.
Figure 3-8 shows the simulated and measured benthic concentrations for Study C. The
data clearly show the distinct influence of the two pesticide applications to the overlying water.
PFAM also simulates the two applications quite well. As expected and as Figure 3-8 shows,
benthic pesticide mass tends to accumulate because of the slow benthic degradation. PFAM
40
-------
also captures the timing of the benthic peaks, which provides support for the default benthic
mass transfer coefficient.
150
10 15
Days After First Application
20
25
Figure 3-7. Simulated water column concentration and measured data for Study C.
O)
£=
0)
o
C
o
o
o
CO
~m
"o
0
C) O O
0
o
o
o
o o
o
o
25
5 10 15 20
Days After First Application
Figure 3-8. Simulated benthic concentration and measured data for Study C.
41
-------
3.2.4 Study D: Texas, Carbaryl Results
Figure 3-9 shows the simulated and reported water levels for Study D. As indicated
earlier, the simulated water level was calibrated by adjusting the weir level three times during the
simulation. The calibration captures the rise and fall of the water level for the most part.
Figure 3-10 shows the water column concentrations and the measured data for Study D.
The two pesticide applications clearly appear in the simulation and the data. The initial
concentration of the first simulated application is similar to the data (about 70% of the
simulation), while the simulated concentration for the second application is about twice that of
the measurements. There was 50% canopy coverage during the application, and crop
interception could be a reason for the discrepancy. In any case, the PFAM simulation is higher
than the data, which is desirable for a regulatory model, and reinforces the appropriateness of the
neglect of the canopy. Again, canopy holdup is an area in need of further research.
Figure 3-11 shows the benthic concentrations for study D. In this case, PFAM simulates
benthic concentrations very well. PFAM captures both the peak and the degradation of the
concentrations in the benthic region. PFAM also captures the lag time for the benthic pesticide
level to reach its peak, providing support for the appropriateness of the assumed value of the
benthic mass transfer coefficient.
PFAM Simulation
Study Data
Figure 3-9. Simulated water level of Study D along with the reported water level.
42
-------
Days
Figure 3-10. Simulated water column concentration and measured data for Study D.
m
O
20
O
80
40 60
Days After Application
Figure 3-11 Simulated benthic concentration and measured data for Study D.
100
43
-------
3.2.5 Study E: California, Triclopyr Results
Study E was a pond study in which the pond level was not regulated but rather allowed to
evaporate without refill. Because the study did not record pond depths during the study, the
hydrodynamics of this study cannot be crosschecked with PFAM output. However, for purposes
of explaining the hydrodynamic effects on pesticide behavior, the water depth simulation is
shown in Figure 3-12. Depth decreases over time during this study, which would tend to buffer
declines in pesticide concentration, but also increase the effective photolysis rate.
Figure 3-13 shows the measured water concentrations along with the PFAM simulation
for Study E, and PFAM captures the data very well. Since trichlopyr degrades predominantly by
photolysis, this study allows a check for the PFAM photolysis routine. The PFAM-simulated
concentration follows a general pattern of decline but degrades somewhat slower than the data
show. This could be due to misestimating the evaporation or could be due to an inexact estimate
of the photolysis rate among other things.
Figure 3-14 shows the measured sediment concentrations along with the PFAM
simulations. PFAM substantially overestimates the sediments concentrations. This
overestimation may be due to the non-site specific fate parameters used for trichlopyr (e.g., the
Kd, benthic degradation, photodegradation rate). Alternatively, the benthic mass could be
influenced by an overestimation of the benthic mass transfer coefficient or a combination of all
of the above. Nevertheless, PFAM errs on the high side, which is acceptable for a regulatory
model.
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0 20 40 60
Days After Application
Figure 3-12. Simulated water depths in the pond for Study E.
80
100
44
-------
. 2.5
D)
£=
O
O
1.5
0.5
O
O
8
0 20 40 60
Days After Application
Figure 3-13. Water column concentrations for Study E.
80
100
D)
O
"ro
o
O
o
CQ
25
20
15
10
0
0
-P-
20
40 60
Days After Application
Figure 3-14. Benthic concentrations for Study E.
80
100
45
-------
3.3 Long-Term Regulatory-Type Evaluation
The following section describes the results of the long-term simulations meant to
represent what regulatory assessors may do with PFAM. A regulatory pesticide assessment will
often address temporal variability that may result from weather by simulating long-term use of a
pesticide with 30 years or so of weather data. Such long-term simulations are often referred to
as Tier 2 assessments. It is also of interest to evaluate the response of PFAM with that of the
currently used simple equilibrium Tier 1 model to identify conditions when PFAM use may be
advantageous. This latter point is the second quality objective of the PFAM validation, that is,
PFAM should provide more refined estimates of concentration over that of the Tier 1 model.
Only the four rice studies (A,B,C, and D) are analyzed here; the pond study (study E) has no
Tier 1 equivalent for comparison and was not included in this long-term analysis.
Figure 3-15 shows the long-term simulation of PFAM along with the current Tier 1
model estimates for the scenario described by Study A. For these long-term simulations, PFAM
was run with the pesticide and water management practices specified in the studies for each year
in the simulations (replicated over 30 years). PFAM estimates in Figure 3-15 include both a
water column concentration as well as the concentration of any released water. Released
concentrations include both intentional releases (i.e., weir is lowered) and overflow releases (i.e.,
excessive rainfall). The Tier 1 concentration in Figure 3-15 is a single value, which represents a
concentration resulting from equilibration with no degradation. PFAM values are substantially
less than the Tier 1 value for both water column and released water. Peak water concentrations
are about 7 ppb, while the maximum released concentration is about 6 ppb as compared with a
Tier 1 estimate of 54 ppb. PFAM chronic concentrations (temporally averaged) are 6 ppb over 7
days and 5 ppb over 30 days, which are also substantially lower than the Tier 1 estimates.
60
50
ra
40
QJ
O
c
o
O
1^
&
03
30
20
10
Tierl
PFAM Water Column
o PFAM Released Water
0 2000 4000 6000 8000 10000 12000
Days
Figure 3-15. Regulatory evaluation using a scenario similar to Study A.
46
-------
Figure 3-16 shows the long-term simulation of PFAM along with the current Tier 1
model estimate for the scenario described by Study B. The maximum PFAM water column
value (80 ppb) exceeds the Tier 1 value (50 ppb), but the maximum released concentration (39
ppb) is well below the Tier 1 value, and released values are the most likely concentrations that
would be used in a risk assessment. The PFAM values exceed the Tier 1 values because the
water level in the simulation dropped below the 10 cm default depth that the Tier 1 model uses.
Therefore, pesticide concentrations in the water column were higher at some times during the
simulation due to the occasional lower amounts of water simulated in PFAM than in the Tier 1
estimate. The PFAM maximum 7-day average (63 ppb) was greater than the Tier 1 value,
whereas the PFAM 30-day average (44 ppb) was slightly less than the Tier 1 value. Released
concentrations were all lower than the Tier 1 values.
on
oU
5
_a.
c
o en
~ bU
1
c
0)
o
O yl fl
o •
(5
I
on
iU
n
-
>
>
<
i
t
i
<
i
<
(
i
j
i:
(
*
(
i
Cl
o
o
c
I
o
c
_J (
-c
>
f
i
Tierl
PFAM Water Columr
PFAM Released Wa
,
i
i
i
I
1
<
'
1
--
(
1
-
1
ter
0 2000 4000 6000 8000 10000 12000
Days
Figure 3-16. Regulatory evaluation using a scenario similar to Study B.
Figure 3-17 shows the long-term simulation of PFAM along with the current Tier 1
model estimate for the scenario described by Study C. In this flow-through scenario, water
continually flows out of the system, so water column and released concentrations are equal.
PFAM values (both released and water column) are less than the Tier 1 value for all times. Peak
water concentrations are about 144 ppb, whereas the Tier 1 estimate is 200 ppb. The 7-day
chronic concentration maximum is 30 ppb and the 30-day maximum average is 14 ppb.
47
-------
250
200
en
a.
-E 150
o
o
O
f
100
50
Tierl
PFAM Column & Released
0 2000 4000 6000 8000 10000 12000
Days
Figure 3-17. Regulatory evaluation using a scenario similar to Study C.
Figure 3-18 shows the long-term simulation of PFAM along with the current Tier 1
model estimate for the scenario described by Study D. The maximum PFAM water column
value (9 ppm) exceeds the Tier 1 value (2.6 ppm), but the maximum released concentration (1.8
ppm) is well below the Tier 1 value. The PFAM values exceed the Tier 1 values because the
water level in the simulation was allowed to drop to around 4 cm, which is well below the 10 cm
that the Tier 1 model uses. Thus the pesticide could become more concentrated in PFAM than in
the Tier 1 simulation. The 7-day chronic concentration maximum is 3.2 ppb, which is above the
Tier 1 value, and the 30-day maximum average is 0.89 ppb, which is below the Tier 1 value.
10
9
cb _,
_§ 7
I 6
ro
o
8 4
QJ o
"5 3
Ti
PF
0 PF
31-1
rAM Water Column
rAM Released Water
ft »
ft-C
t-C
t~a
e
J
I <
)
l_«
D-d
a-t
t — i
s_«
t—t
f
^
C
_ (
> i
> f
^ 4
)
b
2000
4000
Days
6000
8000
Figure 3-18. Regulatory evaluation using a scenario similar to Study D
48
-------
These figures show that PFAM can provide significant refinements of water
concentrations resulting from flooded agriculture as compared to the Tier 1 model. PFAM has
many features not accounted for in the Tier 1 model, such as actual water levels, degradation,
and water management practices. Additionally, PFAM provides estimates for concentrations of
water releases rather than simply a water column concentration, as does the Tier 1 model.
Finally, PFAM considers temporal variability and is able to provide temporal averages that the
Tier 1 model cannot.
3.4 Summary
This work reviewed five aquatic dissipation studies and compared them to PFAM
simulations. PFAM met the two specified quality objectives: 1) PFAM tended to err on the high
side when compared to the actual study data, and 2) PFAM offered more refined estimates than
the Tier 1 model. PFAM estimates, although conservative, were generally within an order of
magnitude of the data but always below the Tier 1 estimates. Thus, PFAM should provide
advantage in the regulation of chemicals for flooded agriculture.
Most of the studies were in reasonable agreement, but there were a few cases, most
notably (Studies B and E), where PFAM estimated the water column concentrations reasonably
well but substantially overestimated the benthic concentrations. The cause of those
discrepancies is difficult to evaluate because of the complex nature of the field studies coupled
with the unavailability of site-specific environmental fate data. Future work may be warranted in
the area of benthic mass transfer coefficient and benthic compartment sizes as these are two
default parameters that could account for some of these discrepancies. Nevertheless, the default
values do provide appropriately conservative estimates with respect to a regulatory model.
When compared with the Tier 1 model, PFAM provided improved estimates that are
closer to field data. Most notably PFAM could account for temporal variability and degradation
that the Tier 1 model does not consider. PFAM water column estimates were generally lower
than the Tier 1 estimates. However, there were cases where the actual water levels in the field
were lower than the preset value in the Tier 1 model, and in those cases, the PFAM estimates
were higher for short periods, as would be expected. In all cases, however, PFAM estimates of
the concentration of releases water were lower than the Tier 1 model, and these released
concentrations are the most relevant concentrations for pesticide risk assessments on non-target
areas.
As previously stated, a regulatory model will be successful if it is protective by not
underestimating the concentrations of measured data and shows reasonable trends in
concentrations in response to the environment. Additionally a higher-tiered regulatory model
should provide advantages over models of lower tiers. PFAM meets these criteria and can
provide appropriate water concentrations for regulatory work involving pesticide use for flooded
agriculture applications.
49
-------
4. User Guidance
4.1.1 Background and Purpose
The goal of PFAM development was to produce a flooded agriculture model for use in
pesticide regulatory work. The user interface was specifically designed with this intent.
Although PFAM can be readily used for analysis of specific sites in a research capacity, its
greatest strength is for analysis of long-term pesticide use for hypothetical scenarios, as is
typically done during the pesticide registration process.
As is typical in a regulatory pesticide assessment, model simulations are performed over
a long period with pesticide applications occurring every year. The PFAM user interface is
designed for this application. The duration of PFAM simulations will be as long as the available
meteorological data (typical available files are 30 years). Pesticide application and flooding
sequences are mapped onto the time series in 1-year cycles for as long as the simulation
continues. Thus, as is typical of a U.S. exposure assessment, the pesticide is assumed to be used
every year for 30 years (the length of the U.S. meteorological files). Currently output is
delivered as released mass of pesticide as well as a daily time series of concentrations.
4.1.2 Quick Start
The installation package comes with a test input file Test.pfa and an example
meteorological file wTest.dvf. After start up, use File/Retrieve on the upper left menu bar to load
the example file (Test.pfa). The user will need to specify where the output is delivered and
should go to the Output tab and specify the desired location and name. The user will also need to
specify the location of the meteorological data and should go to the Location tab and specify the
file (the wTest.dvf file in the program directory is available as an example meteorological file).
Press the Run button and output should be delivered in two files: raw data into the file previously
specified as the output and processed data in a similarly named file but appended with .ppl.
(There will also be receiving water body output if that option was selected).
4.1.3 Menu Items
File manipulations are performed on the menu bar. The first menu item is File., with
submenus Retrieve Input and Save Input. Retrieve Input will open up a file browser dialog and
allow a user to upload a previously created input file into the PFAM interface. The input files
are text files that can be created either with the PFAM interface or with a text editor. The Save
Input command will open a file browser and allow the user to save the inputs from the PFAM
interface into a text file.
4.1.4 Chemical Tab Sheet
The chemical properties tab sheet allows entering of the chemical properties of the
pesticide. The definitions are typically those generally accepted by the scientific community and
are summarized here:
Water Column Half-Life [d] is the half-life of a pesticide resulting from metabolic processes.
This parameter acts on all phases of the pesticide in the water column. As is typical for the type
of metabolism studies submitted for pesticide registration, no distinction is possible between
sorbed-phase degradation and aqueous-phase degradation, and only total system degradation is
50
-------
known. Thus no distinction is made for this application either. A half-life of zero is interpreted
to mean that the compound does not degrade by this process.
Temperature Associated with the Water Column Value [°C] is the temperature at which the
water column degradation study was conducted. Degradation rate is adjusted by temperature
with this temperature input being the reference.
Benthic Compartment Half-Life [d] is the half-life of a pesticide in the flooded benthic
compartment. Typically, this value could be an anaerobic metabolic half-life. This parameter
acts on all phases of the pesticide in the benthic compartment. As is typical for the type of
metabolism studies submitted for pesticide registration, no distinction is possible between
sorbed-phase degradation and aqueous-phase degradation, and only total system degradation is
known. Thus no distinction is made for this application either. A half-life of zero is interpreted
to mean that the compound does not degrade by this process.
Temperature Associated with the Benthic Compartment Value [°C] is the temperature at which
the anaerobic metabolism study was conducted.
Unflooded Soil Half-Life [d] is the half-life of a pesticide on the soil in unflooded conditions.
This half-life acts on the "benthic" compartment when that compartment is not flooded. This
value should include all expected mechanisms of degradation in the soil compartment, including
metabolism, volatilization, and photolysis. This parameter acts on all phases of the pesticide in
the benthic compartment. As is typical for the type of metabolism studies submitted for pesticide
registration, no distinction is possible between sorbed-phase degradation and aqueous-phase
degradation, and only total system degradation is known. Thus no distinction is made for this
application either. A half-life of zero is interpreted to mean that the compound does not degrade
by this process.
Note that the routines to account for photolysis and volatilization are turned off when the field is
not flooded and users must manually enter the effective half-life to include photolysis or
volatilization on dry soil. The photolysis and volatilization routines are turned back on when the
field floods. The reason for this behavior is that the photolysis and volatilization routines were
written specifically for aquatic systems, not dry fields. Furthermore, OPP has not yet accepted a
field volatilization or photolysis routine for any of its models (e.g., PRZM) for use in risk
assessments; therefore those processes were left out until approval by USEPA OPP.
Temperature Associated with the Unflooded Soil Value [°C] is the temperature at which the soil
degradation study was conducted.
Photolysis Half-Life [d] is the near-surface 24-hr average aquatic half-life of the pesticide due to
photolysis. A half-life of zero is interpreted to mean that the compound does not degrade by this
process.
Photolysis Reference Latitude [°] is the latitude that the photolysis value is intended to simulate.
51
-------
Hydrolysis Half-Life [d] is the half-life of the pesticide due to hydrolysis at whatever pH is to be
simulated. A half-life of zero is interpreted to mean that the compound does not degrade by this
process.
Koc [mL/g] is the organic-carbon-normalized sorption coefficient. In some situations, use of Koc
may not be appropriate, and a Kd value may be preferred. Kd can be applied by making note of
the oc content (Foe Benthic under Physical Tab) and recognizing that Kd = Koc x oc, and
adjusting the Koc as appropriate.
Molecular Weight [g/mol] is the molecular weight of the pesticide. Molecular Weight is used
only in the volatilization routine. This parameter only affects the volatilization rate.
Vapor Pressure [torr] is the vapor pressure of the compound at a representative temperature to
be simulated. This parameter only affects the volatilization rate.
Solubility [mg/L] is the solubility of the pesticide at a representative temperature to be simulated.
Solubility is used only in the volatilization routine; it does not cap concentrations in this
program. This parameter only affects the volatilization rate.
Heat of Henry [-] is the enthalpy of phase change from aqueous solution to air solution
(Joules/mole). This enthalpy can be approximated from the enthalpy of vaporization
(Schwarzenbach et al., 1993), which can be obtained from EPISuite among other sources.
Enthalpy for pesticides obtained in a literature review ranged from 20,000 to 100,000 J/mol
(average 59,000 J/mol). Some example enthalpies for pesticides are
Metalochlor 84,000 Feigenbrugel et al. 2004
Diazonon 98,000 Feigenbrugel et al. 2004
Alachlor 76,000 Gautier et al., 2003
Dichlorvos 95,000 Gautier et al., 2003
Mirex 91,000 Yin and Hassett, 1986
Lindane 43,000 Staudinger et al. (2000)
EPTC 37,000 Staudinger et al. (2000)
Molinate 58,000 Staudinger et al. (2000)
Chlorpyrifos 17,000 Staudinger et al. (2000)
Enthalpies can also be estimated by the US EPA EPI Suite software. Open the software, then
select the FIENRYWIN subprogram on the left of the EPI Suite screen. On the top menu of the
HENRYWIN window item, select the ShowOptions, then select Show Temperature Variation
with Results. Enter the chemical name of interest and then push the Calculate button. EPI Suite
will give the temperature variation results in the form of an equation: HLC (atm-m3/mole) =
exp(A-(B/T)) (T in K}. The enthalpy of solvation in Joules/mol is equal to 8.314*B. Example
enthalpies from EPI Suite are:
Pendamethalin62,000 J/mol
Carbaryl 58,000 J/mol
Carbofuran 54,000 J/mol
52
-------
Molinate 54,000 J/mol
Endosulfan 37,000 J/mol
Henry Reference Temperature [°C] is the temperature at which the vapor pressure, solubility,
and Henry's Law constant apply.
4.1.5 Applications Tab Sheet
The Applications tab refers to pesticide applications and contains a table that allows entry
of a number of pesticide applications during a growing season. Each application is identified by
the date (month and day) at which it is applied and the amount of pesticide that is applied
(kg/hA).
Number of Applications. Enter the number of applications and press the Update button. This
will make available only the number of applications requested. (Information is retained in the
invisible text boxes, but it is not used.).
Day is the entry for the calendar day (1 through 31) of pesticide application.
Mon is the calendar month (1 through 12, corresponding to January to December) of pesticide
application.
Applied Mass [kg/ha] is the mass per area at which the pesticide is applied on the associated
date.
Slow Release [d"1] specifies whether the pesticide is designed as a slow-release agent. The
release rate is first order with the mass of pesticide remaining after time t equal to Moexp(-kt),
where k is the user specified release rate (per day), Mo is the Applied Mass (see above), and t is
time in days. A slow release of zero by convention means that there is an instantaneous full
release. As a note of reference, a Slow Release rate of 0.6 per day will result in 95% of the
pesticide released in 5 days.
4.1.6 Location Tab Sheet
The Location sheet allows specification of particular weather files. The weather file
should be organized without a header and into the following white-space-delimited columns:
date, precipitation (cm), pan evaporation (cm), average temp (°C), wind speed (cm/s).
The date should be presented as a number consisting of the two digit numerical values for
month day year and compiled together. For example December 15, 1992 should be written as
121592, and January 3, 1991 should be written as 010391 (or 10391). The program will read
the entire date value in as a single integer and parse the value. Because the file is recognized as
being white-space delimited, the date should not contain any internal spaces. For example,
February 7, 1992 which is 020792 can be written as 20792 but not as 2 792.
Daily meteorological files for the United Sates that will work for PFAM are available
from the US EPA at: http://www.epa.gov/ceampubl/tools/metdata/index.htm. The files at that
address contain additional columns of information that have no effect on PFAM.
53
-------
4.1.7 Flood Events Tab Sheet
The Flood Events sheet allows specification of up to 11 flood events, in which an event
refers to a manually controlled change in water level. Flood events are mapped onto the time
series every year with the mapping unit defined as Event 1 to Event 1 of the next year and does
not depend on the calendar year. Thus flood event series can be defined across calendar years if
so desired. The events should be listed in chronological order and no event should occur past
365 days. There does not have to be any actual significance to the Event 1 date. It could simply
be January 1 and then all events would be referenced to the first of the year. Or, it could be the
actual day of the first flood event, and then all subsequent events would be referenced to the first
flood. However, at the start of the simulation weir height, fill height, minimum level, and
turnover are initialized to 0 and remain that way until the first event. Thus, if it is desired to have
other than zeroes for these values at the start of the simulation, then Event 1 should be defined as
the earliest event relative to the start of the simulation (in cases for USEPA dvf files this would
be January 1).
Event 1 Day is an arbitrary calendar day that when used along with the Event 1 Month
represents a reference point for specifying flood events. This should be entered as an integer
calendar day with an acceptable value range dependent upon the month selected.
Event 1 Month is an arbitrary calendar month that when used along with the Event 1 Day will be
used as a reference point for specifying flood events. This should be an integer calendar month
(1 through 12, corresponding to January to December).
Number of Days represents the number of days since the Event 1 date when a respective event
occurs. This should be entered as an integer. (Event 1 is always zero.)
Weir Level [m] represents the level that the weir was set to on the respective event. This level is
the level at which overflow occurs.
Fill Level [m] is the level that the water is filled to for a refilling event. If a Fill Level value is
higher than the previous Fill Level value, then it is assumed that the user wants to increase the
water level manually, and on that day, the water level will be increased to the Fill Level.
(Reductions in the Fill Level value do not have any immediate effect in a likewise manner, but it
will affect the subsequent refill events.)
Minimum Level [m] represents the water level that the water body can be reduced to (by
evaporation) before the level is returned to the weir level by refill.
Turn Over [d"1] represents the turnover rate for each event. It is determined by dividing the flow
through the water body by the volume of the water body. Turnover is used for those systems
that use a continuous flow through the system and are maintained at a set depth. (Turnover
should be set to zero when weir level is zero.)
4.1.8 Crop Tab Sheet
The Crop sheet allows specification of crop growth characteristics.
54
-------
Zero Height Reference Day represents the day of the month when a crop just emerges (i.e., zero
height). It is used along with a Zero Height Reference Month.
Zero Height Reference Month represents the month when a crop just emerges (i.e.., zero height).
It is used along with a Zero Height Reference Day.
Days from Zero to Full Height represents the number of days between the zero reference day
and the time that the plant reaches full height.
Days from Zero to Removal represents the number of days between the zero reference day and
the time that the plant is removed.
Maximum Fraction Areal Coverage represents the maximum fractional area that the plant
covers at full height.
4.1.9 Physical Tab Sheet
The Physical sheet allows manipulation of parameters that characterize the physics of the
water body.
Mass Transfer Coefficient [m/s] represents the mass transfer coefficient between the water
column and the benthic zone. It accounts for all means of mass transport and is referenced to the
surrogate driving force of aqueous concentration differences. It is a difficult parameter to
measure. Literature and EPA's own calibrations suggest a starting estimate of 10~8 m/s.
Leakage [m/d] is the rate of water flowing from the water column through the benthic layer and
out of the bottom of the system. Leakage will only occur if there is water in the water column.
Reference Depth [m] represents the depth that aquatic measurements were made when
determining factors such as suspended solids, biota, and DOC [m]. This usually will be set to the
typical depth of the water body.
Benthic Depth [m] is the depth of the benthic compartment. This is another difficult to estimate
parameter; however, literature and EPA's own calibrations suggest about 0.05 m.
Benthic Porosity is the porosity of the benthic compartment: [pore space volume per total
volume]. The default value is set to that used by the EPA standard farm pond.
Dry Bulk Density [g/cm3] is the rationally defined bulk density: [mass of sediment per total
volume of sediment]. The default value is set to that used by the EPA standard farm pond.
Foe Water Column SS is the fraction of organic carbon associated with suspended sediment.
The default values are set to those used by the EPA standard farm pond.
Foe Benthic is the fraction of organic carbon associated with benthic sediment. The default
value is set to that used by the EPA standard farm pond.
55
-------
SS [mg/L] is the suspended mass in the water column. The default value is set to that used by the
EPA standard farm pond.
CHL [mg/L] represents the chlorophyll concentration in the water column. The default value is
set to that used by the EPA standard farm pond. This parameter only affects the photolysis rate.
Water Column DOC [mg/L] represents the dissolved organic carbon concentration in the water
column. The default value is set to that used by the EPA standard farm pond. This parameter
only affects the photolysis rate.
QT [-] is the Qio value for metabolism. A value of 2 is typical for such models.
DFAC [-] is a parameter defined as is in the model EXAMS. It represents the ratio of vertical
path lengths to depth. Default value is set tol.19 as suggested by EXAMS documentation.
4.1.10 Tab Sheet for Output
The Output sheet allows specification of a file and a directory that will contain output.
The Select Output File button opens a file browser and allows a user to select or create an output
file name and location. A raw data file will be created with the user-specified name. An
additional file will also be created that contains post-processed results derived from the raw data
file. The post processed file will have the same name as the raw data file but will be appended
with ".ppl". Users may produce their own post processor and replace the post processor if so
desired. The post processor is pfamppl.exe. The command line call that PFAM uses internally
to call the post processor is the namepfamppl.exe and a command line argument that contains
the raw data file location in quotation marks, that is,
>pfamppl.exe "full path and file name of raw data file"
Additionally the program will also check for the existence of a second post processor that a user
may create. If a file exists with the name pfampp2.exe, then the program will execute the
following command:
>pfampp2.exe "full path and file name of raw data file"
Area of Application [m2] Area where the applications are made (field or paddy area)
Route effluent check box will start a post processor that will route the effluent from PFAM into
EPA standard water bodies (pond and reservoir). This parameter also routes a surrounding field
runoff into the receiving body.
Area of Surrounding Watershed [m2] Area of the watershed excluding the application areas.
This value along with the curve number will determine the amount of runoff the receiving water
body gets from non-application areas.
Curve Number of Surrounding Field is the Natural Resources Conservation Service (NRCS)
curve number of the surrounding watershed. This will determine the amount of runoff the
receiving body gets from non-application areas.
56
-------
Width of Mixing Cell [m] is the width of the receiving water body.
Depth of Mixing Cell [m] is the depth of the receiving water body.
Length of Mixing Cell [m] is the length of the receiving water body. Note that for a flowing
water body such as a stream or river, this length value should correspond roughly to twice the
dispersivity (2D/v, where D is the dispersion coefficient and v is the velocity of the stream or
river) characteristic of the flowing water body. A good starting value may be around 30 meters
as estimated from the median of data in Fisher et al. (1979).
Base Flow [m3/sec] is the base flow through the receiving water body.
See Current Output and Post Processing at the end of this manual for details on the currently
supplied post processing.
4.1.11 Degradate 1
The Degradate 1 tab sheet is used if a degradate is to be produced from degradation of
the main chemical. If degradate calculations are desired, then check the Perform Degradate
Calculation checkbox. Chemical properties for Degradate 1 are defined in the same manner as
those described on the Chemical tab sheet for the parent chemical, except with the additional
entries as follows:
Moles of Degradate Produced per Mole of Parent. These entries indicate how many moles of
degradate are produced by the degradation of the parent for each of the specific processes.
4.1.12 Degradate 2
The Degradate 2 tab sheet is similar to the Degradate 1 sheet except that the molar productions
of Degradate 2 are referenced to Degradate 1.
4.1.13 Run Button
The run button gathers all input from the controls in the PFAM interface. The program
performs some checking of the values and then runs PFAM. Output is produced as described
above.
Current Output and Post Processing
Assuming that the output file is named output.txt, then the following files will be
generated:
Output.txt: raw daily output data from PFAM
Output_Efffective_Halflife.txt: summary of the effective average dissipation processes for the
PFAM simulation
Output_ProcessedOutput.txt: summary of PFAM releases and the reason for the releases. Also
gives the daily concentrations in the water column and the benthic/soil compartment.
57
-------
0wrpw^_ReceivingBodies.txt: water column results for receiving body. Result of post processing
the PFAM output by routing the PFAM effluent into the standard pond and the standard reservoir
and through a flowing water body (mixing cell).
0wZpw^_ReceivingBodies_daily.txt: daily water column concentrations in the receiving water
bodies
0wZpw^_ReceivingBodies_Benthic.txt: benthic pore water results for receiving body. Result of
post processing the PFAM output by routing the PFAM effluent into the standard pond and the
standard reservoir and through a flowing water body (mixing cell).
0wZpw^_ReceivingBodies_daily_Benthic.txt: daily benthic pore water concentrations in the
receiving water bodies
5. References
Banks, R. B., 1975. Some Features of Wind Action on Shallow Lakes. Journal of the
Environmental Engineer ing Division., ASCE. 101(EE5), 813-827.
Burns, L.A., Suarez, L.A., Prieto, L.M., 2007. United States Meteorological Data, Daily and
Hourly Files to Support Predictive Exposure Modeling. United States Environmental Protection
Agency, Washington DC. EPA/600/R-07/053 May 2007
Burns, L.A., 2000. Exposure Analysis Modeling System (EXAMS): User Manual and System
Documentation. EPA/600/R-00/81-023, U.S. EPA.
Crout, N.M.J., Tarsitano, D., and Wood, A.T., 2009. Is my model too complex? Evaluating
model formulation using model reduction. Environmental Modelling & Software. 24(1), 1-7.
Cusler, E.L., 1984. Diffusion: Mass Transfer in Fluid Systems, Cambridge University Press, New
York.
Feigenbrugel,V., Calve,S.L., Mirabel,P., Louis, F. (2004). Henry's law constant measurements
for phenol, o-, m-, and p-cresol as a function of temperature, Atmospheric Environment, 38(33),
5577-5588
Fisher, H.B, List, EJ, Koh, R.C.Y., Imberger, J., and Brooks, N.H. (1979) Mixing in Inland and
Coastal Waters. Academic Press, New York, NY. pp. 126-127.
Freni, G., Mannina, G., Vivianni, G., 2009. Identifiability analysis for receiving water body
quality modelling. Environmental Modelling & Software. 24(1), 54-62.
58
-------
Gautier, C., Stephane Le Calve, Philippe Mirabel (2003). Henry's law constants measurements
of alachlor and dichlorvos between 283 and 298 K, Atmospheric Environment, 37(17) 2347-
2353.
Hamon, W.R. 1961. Estimating Potential Evapotranspiration, Journal of the Hydraulics Division,
ASCE87(HY3):107-UO
Jeon, J.H., Yoon, C.G., Ham, J.H., Jung, K.W., 2005. Model Development for Nutrient Loading
Estimates from Paddy Rice Fields in Korea. Journal of Environmental Science and Health, Part
B, 39(5-6) 845-860.
Jeon, J.H., Yoon, C.G., Donigan, A.S., and Jung, K.W., 2007. Development of the HSPF-Paddy
model to estimate watershed pollutant loads in paddy farming regions. Agriculture Water
Management, 90(1-2), 75-86.
Johnson, B.R., 1991. A Simple Adsorption/Dilution Model for Rice Herbicides. Bull. Environ.
Contam. Toxicol., 47(2), 244-250.
Karpouzas, D.G. and Capri, E. 2006. Risk Analysis of pesticide applied to rice paddies using
RICEWQ 1.6.2v and RIVWQ2.02. Paddy Water Environment, 4(1), 29-38.
Khepar, S.D., Yadov, A.K., Sondhi, S.K., and Siag, M. 2000. Water Balance Model for Paddy
Fields under Intermittent Irrigation Practices. Irrig Sci., 19(4), 199-208.
Liss, P.S., 1973. Processes of Gas Exchange Across an Air-Water Interface. Deep Sea Research,
20(3), 221-238.
Luo, Y. 2011. Report 263: Review and Evaluation of Pesticide Modeling Approaches in Rice
Paddies, Department of Pesticide Regulations, Sacramento, CA.
NRC, 2007. Models in Environmental Regulatory Decision Making. National Research Council,
The National Academies Press, Washington, DC.
NRCS, 2003. National Engineering Handbook. Section 4: Hydrology. National Soil
Conservation Service, USDA, Washington, DC.
Petty D.G., Skogerboe, J.G., Getsinger, K.D., Foster, D.R., Houtman, B.A., Fairchild, J.F.,
Anderson, L.A., (2001). The Aquatic Fate of Triclopyr in Whole-Pond Treatments. Pest
Management Science, 57(9), 764-774.
Ranatunga, K., Nation, E.R., and Barratt, D.G., 2008, Review of soil water models and their
applications in Australia. Environmental Modelling & Software. 23(9), 1182-1206.
Rykiel, E. 1996. Testing Ecological Models: The Meaning of Validation. Ecological Modeling
90(3), 229-244.
59
-------
Shwarzenbach, R.P., Gschwend, P.M., and Imboden, D.M. 1993. Environmental Organic
Chemistry, John Wiley & Sons, New York.
Staudinger, J. and Roberts, P.V., 2001. A Critical Compilation of Henry's Law Constant
Temperature Dependence for Organic Compounds in Dilute Aqueous Solutions. Chemosphere,
44(4), 561-576.
Tournebize, J., Watanabe, H., Takagi, K., Nishimura, T. 2006. The development of a coupled
model (PCPF-SWMS) to simulate water flow and pollutant transport in Japanese paddy fields.
Paddy Water Environment, 4(1), 39-51.
USEPA, 1992. Data submission, Master Record Identification Number 42560502.
USEPA 1994. Data Submission, Master Record Identification Number 43263001.
USEPA, 1997a. Data Submission, Master Record Identification Number 44456103.
USEPA, 1997b. Reregi strati on Eligibility Decision (RED) for Pendimethalin. United States
Environmental Protection Agency, EPA 738R-97-007, June 1997.
USEPA, 1998a. Data Submission, Master Record Identification Number 44527601.
USEPA, 1998b. Reregi strati on Eligibility Decision (RED) for Triclopyr. United States
Environmental Protection Agency, EPA 738R-98-011, June 1997.
USEPA, 1999. Data Submission, Master Record Identification Number 4889236.
USEPA, 2001. August 06, 2001. Memorandum. 58 Page(s). Thomas Steeger. Environmental
Risk Branch. EFED Risk Assessment for Section 3 Registration of Bispyribac-Sodium
(Sodium 2,6-bis[(4,6-dimethoxypyrimidin- 2yl)]benzoate)(Regiment). In USEPA Electronic
Reading Room: Science Reviews Previously Released Under FOIA
(http://www.epa.gov/pesticides/foia/reviews.htm)
USEPA, 2002. Guidance for Quality Assurance Project Plans for Modeling, EPA QA/G-5M. US
Environmental Protection Agency, EPA/240/R-02/007. Washington DC.
USEPA, 2004. Refined (Level II) Terrestrial and Aquatic Models for Probabilistic and
Ecological Assessments of Pesticides. March 30 -April 2, 2004. Docket Number: OPP-2004-
005-002
USEPA, 2006. Reregi strati on Eligibility Decision (RED) for Propiconazole. United States
Environmental Protection Agency, EPA 738R-06-027, July 2006.
USEPA 2007a, Tier I Rice Model vl.O for Estimating Pesticide Concentrations in Rice Paddies.
Available at http://www.epa.gov/oppefedl/models/water/rice tier i.htm.
60
-------
USEPA 2007b. Carbaryl Refined Drinking Water Time Series Simulations Using PC As March
9, 2007. Docket ID EPA-HQ-OPP-2007-0941-0014
USEPA, 2009a. Guidance on the Development, Evaluation, and Application of Environmental
Models. EPA/100/K-09/003 March 2009. US Environmental Protection Agency, Washington,
DC
USEPA, 2009b. Estimations Program Interface Suite, v 4.0, United States Environmental
Protection Agency, Washington, DC.
Vanderborght, J.P. and Wollast, R. 1977. Mass Transfer Properties in Sediments near Benthic
Layers. Bottom Turbulence., J.CJ Nihoul ed., Elsevier Scientific Publishing Co., New York.
209-219.
Watanabe, H. and Takagi, K. 2000. A Simulation Model for Predicting Pesticide Concentrations
in Paddy Water and Surface Soil. I. Model Development. Environmental Technology, 21 1379-
1391.
Yin, C.; Hassett, J. P. (1986) Gas-partitioning approach for laboratory and field studies of mirex
fugacity in water. Environ. Sci. Technol. 1986, 72, 1213-1217.
Yoshinaga, I, Feng, Y., Singh, R.K. and Shiratani, E. 2004. Dissolved Nitrogen Model for
Paddy Field Ponded Water During Irrigation Period. Journal Paddy and Water Environment
2(3), 145-152.
Young, D.F., 2012. Development and Evaluation of a Regulatory Model for Pesticides in
Flooded Applications. Environmental Modeling & Assessment 17(5), 515-525. DOI:
10.1007/sl0666-012-9311-9.
61
------- |