United States
Environmental Protection
Agency
Research and
Development
Assessing Nitrogen Leaching
in Unsaturated Media from
Septic Tank Waste for the
North Animas Valley, La Plata
County, Colorado
A Wellhead Protection
Screening Assessment
and Planning Document
Ecosystems Research Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Athens, Georgia
September 1996
-------
Assessing Nitrogen Leaching in Unsaturated Media from Septic Tank Waste for the North
Animas Valley, La Plata County, Colorado: A Wellhead
Protection Screening Assessment and Planning Document
by
Robert F. Carsel
Regulatory Support Branch
Ecosystems Research Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Athens, GA 30605-2720
-------
Section 1
Background and Purpose:
The U.S. Environmental Protection Agency is continually faced with regulatory
issues concerning the migration of organic and inorganic chemical constituents to and
through groundwater systems. Each of these issues requires that the potential risk to
human health resulting from the introduction or continued use of such chemicals be
evaluated. Recently, much of this attention has been focused on exposure through leaching
of nitrogen (specifically nitrates) to groundwater and subsequent ingestion of contaminated
water. Historically, nitrates have been associated with application in an agricultural crop
production setting; however, nitrates can also be transported to groundwater from septic
tanks at residences not served by a municpal waste treatment system.
The capability to simulate the potential exposure to nitrates via aricultural or
residential seepage pathways has two major facets:
i
ISF Prediction of the fate of the chemical, after it is applied or discharged, as it is
transported by water down through the soil into groundwater.
ra" Evaluation of the probability of occurrence of concentrations of various
magnitudes at various depths.
Several models are capable of simulating the transport and transformation of
nitrogen in the soil subsurface. None of these models, however; have been linked together
to provide a complete simulation package of unsaturated zone models that have the
flexibility to handle a wide variety of hydrogeologic, soils, climate, and source control
scenarios. The formulation of the risk analysis problem, however, requires more than a
simple, deterministic evaluation of potential exposure concentrations. The inherent
variability of forcing functions, storage and retardation terms in natural systems, combined
with the inability to describe these attributes of the system exactly, suggest that exposure
concentrations cannot be predicted with certainty. Therefore, the uncertainty associated
with the predictions must be quantified. Consequently, this simulation package also seeks
to provide uncertainty analyses through Monte Carlo simulation techniques.
Exposure assessment, as defined for human impacts, is the estimation of the
magnitude, frequency, and duration at which a quantity of a toxicant is available at certain
exchange boundaries (i.e., the lungs, gut or skin) of a subject population over a specified
time interval. The concentration estimates generated during an exposure assessment are
combined with demographic and toxicological information to evaluate risk to a population
which can be used, in turn, to make policy decisions regarding the use and disposal of
chemicals.
The 1986 Amendments to the Safe Drinking Water Act require each state to develop
and submit to the U.S. EPA a wellhead protection (WHP) program. As part of the program,
states must establish procedures for delineating wellhead protection areas (WHPA) around
-------
each water well or well field that supplies a public potable water system. A wellhead
protection area (WHPA) is defined as the surface and subsurface area surrounding a water
well or well field through which contaminants are likely to be transported and reach the
well or wellfield. Within the WHPA, contaminant sources need to be assessed and managed
to prevent pollution of public drinking water supplies. Existing WHP programs are
generally aimed at one of three overall protection goals:
Provide a remedial action zone to protect wells from unexpected contaminant
releases.
Provide an attenuation zone to bring concentrations of specific contaminants
to desired levels at the time they reach the wellhead.
Provide a well-field management zone in all or part of a well's present or
future recharge area.
Numerical models provide great flexibility and accuracy in representing complex
environments and can be applied to nearly all types of hydrogeologic settings. The models
can also be used to predict the dynamic aspects of the WHPA, such as changes in the size of
the WHPA resulting from natural or man-made effects. Disadvantages for this method
include costs that are high relative to other methods and the need for considerable technical
expertise in hydrogeology and modeling. The cost may be warranted in areas where a high
degree of accuracy is desired, however. Also, due to limitations on model grid spacing and
density, numerical models are sometimes less suitable than analytical methods for
assessing drawdowns close to pumping wells.
The more rigorous the method used for WHPA delineation, the smaller the WHPA
can be without risking underprotection and the associated potential for water quality
degradation. When a smaller WHPA can be defined without generating unacceptable risk,
land use restrictions can be kept to a minimum along with the potential economic hardships
associated with those land use restrictions. The choice of WHPA delineation methodology
becomes a decision based on generating an acceptable margin of safety, while balancing the
economic hardships to affected parties with the technical and economic feasibility of
minimizing the WHPA.
Nationwide, housing areas are located in many diverse hydrogeologic environments
(e.g., multiple aquifer systems, fractured and/or karst systems, and systems with wide
variations in depth to the water table). In addition, recharge can vary widely because of
irrigation practices and/or climate. Also, domestic and irrigation wells, which pump at
different and varying rates, are commonly located throughout agricultural regions.
Therefore, the ability to model transient flow conditions (i.e., transient recharge, a
fluctuating water table, and transient pumping from a variety of points in x,y,z space) for a
wide variety of hydrogeologic conditions is important.
Contamination scenarios in housing regions must consider multiple point and
nonpoint source loadings that vary both spatially and temporally. For example, each home
in a housing complex can have a septic tank which can result in point sources of
-------
contamination to aquifers.
Nitrogen loadings to the subsurface are affected by surface processes and source
control (e.g., limitations on housing density) management practices. Most of these processes
require detailed modeling of subsurface flow and transport. Text or matrix ranking or the
separate application of an existing model frequently is used to estimate recharge and solute
loading to the subsurface.
The contaminants of concern in a housing development having septic tanks are
predominantly nitrates. Other forms of nitrogen, including nitrite and ammonia, may also
be present in the subsurface at varying concentrations. Because interest in housing areas is
likely to focus on dilute forms of nitrogen, issues such as the source of waste into the septic
tank, processes within the septic tank and discharge of nitrogen out of the septic tank are
critical to the overall loading to groundwater. The ability to provide screening assessments
identifying which parts of either an existing or proposed new housing development may
contribute the highest leaching of nitrates is essential as a first step to provide wellhead
protection strategies.
1.0 Objectives
The Pesticide Root Zone Model (PRZM-2) is a one-dimensional, dynamic,
compartmented model that can be used to simulate chemical movement in unsaturated soil
systems within and immediately below the plant root zone. It has two major components-
hydrology (and hydraulics) and chemical transport. The hydrologic component for
calculating runoff and erosion is based on the Soil Conservation Service curve number
technique and the Universal Soil Loss Equation. Evapotranspiration is estimated either
directly from pan evaporation data or indirectly using an empirical formula.
Evapotranspiration is divided among evaporation from the plant canopy interception,
evaporation from soil, and transpiration by the plant. Water movement is simulated by the
use of generalized soil parameters, including field capacity, wilting point, and saturation
water content. The chemical transport component can simulate pesticide application on the
soil or on the plant foliage. With a newly added feature, biodegradation can also be
considered in the root zone. Dissolved, adsorbed, and vapor-phase concentrations in the soil
are estimated by simultaneously considering the processes of pesticide uptake by plants,
surface runoff, erosion, decay, volatilization, foliar washoff, advection, dispersion, and
retardation. Two options are available to^ solve the transport equations: (1) the original
backwards-difference implicit scheme that may be affected by excessive numerical
dispersion at high Peclet numbers, or (2) the method of characteristics algorithm that
eliminates numerical dispersion while slightly increasing model execution time.
PRZM-2 has the capability to simulate multiple zones. This allows PRZM-2 and
VADOFT (the vadose zone model) to combine different root zone and vadose zone
characteristics into a single simulation. Zones can be visualized as multiple land segments
joined together in a horizontal manner.
-------
There are three reasons a user may choose for implementing multiple zones:
1) to simulate heterogeneous PRZM root zones with a homogeneous vadose zone
2) to simulate a homogeneous root zone with heterogeneous vadose zones
3) to simulate multiple homogeneous root zones with multiple homogeneous
vadose zones.
Another added feature is the ability to simulate as many as three chemicals simulta-
neously, as separate compounds or as a parent-daughter relationship. This gives the user
the option to observe the behavior of multiple chemicals without making additional runs, or
the ability to analyze the mass transformation and transport behavior of a parent chemical
and one or two of its daughter products. In this application, we simply replace pesticide
chemical properties and transformation pathways with those of nitrogen forms of interest.
Predictions are made on a daily basis. Output can be summarized for a daily,
monthly, or annual period. Daily time series values of various fluxes or storages can be
written to sequential files during program execution for subsequent analysis.
VADOFT is a finite-element code for simulating moisture movement and solute
transport in the vadose zone. It is the second part of the two-component PRZM-2 model for
predicting the movement of pesticides within and below the plant root zone, and assessing
subsequent groundwater contamination. The VADOFT code simulates one-dimensional,
single-phase moisture and solute transport in unconfined, variably saturated porous media.
Fate and transport processes include hydrodynamic dispersion, advection, linear equili-
brium sorption, and first-order decay. The code predicts infiltration, or recharge rate, and
solute mass flux entering the saturated zone.
PRZM-2 can be run in a Monte Carlo mode so that probabilistic estimates of
chemical solute loadings to the saturated zone from the source area can be made. The input
preprocessor allows the user to select distributions for key parameters from a variety of
distributions; the Johnson family (which includes the normal and lognormal), uniform,
exponential and empirical. If the user selects distributions from the Johnson family, he or
she may also specify correlations between the input parameters. The Monte Carlo processor
reads the standard, deterministic input data sets for each model, then reads a Monte Carlo
input file that specifies which parameters are to be allowed to vary, their distributions, the
distribution parameters, and correlation matrix. The model then executes a prespecified
number of runs.
The output processor is capable of preparing statistics of the specified output
variables including mean, maximum values and quantiles of the output distribution. The
output processor also can tabulate cumulative frequency histograms of the output variables
and send them to a line printer for plotting.
Although PRZM-2 (Mullins et al., 1993) has all the basic components necessary for
simulating general chemical migration in unsaturated media and evaluating uncertainty,
-------
the original model does not have the specific capability to model nitrogen species. Use of the
model as a component of a methodology for delineating rural wellhead protection areas
relevant to nitrogen species (in particular, nitrate) requires that the model be enhanced to
represent (1) nitrogen introduced as a result of on-site wastewater treatment systems, (2)
soil nitrogen processes within the unsaturated zone, and (3) certain potential influxes of
nitrogen due to land surface activities related to agriculture and atmospheric deposition.
The objective of this work is to develop a nitrogen modeling system capable of assessing the
impact of septic tank effluent that can be used as an initial screening tool for wellhead
protection. These enhancements are illustrated in Figure 1.0
-------
i j f^iuui . i in t_/vjt_,i i IIC4IIOIWIIHCILIWIIO I
vjy Old I Id
4-Panon RejkJence
180 gal/day with
67% N as Organic
32% N as Ammonium
DnMMd
,7)%ammonlum-N. 19% organic N, 1%nRrat»-N
Much of the Organic N
Converted to Ammonium
IJ | I I I I I I I
10Sof Neventuaty
pumped out as septage
TbW Ntregw h Elfluwrt SO ppm
rapW Convention of ammonium MD nltrat*
1 PPfn
(natural racharo*)
4 J ppm nKiHi>t 1 ppn) amnwoium
-------
1.2 REPORT ORGANIZATION
The next four sections of this document describe the major components of the
nitrogen- enhanced PRZM/VADOFT modeling system that we now designate as PRZM-3. In
Section 2 we focus on soil nitrogen process modeling. First, we present a summary of a
cursory literature review that determined our selection of a modeling scheme for nitrogen
inflows, species, and processes. Following this summary, the soil process algorithms for
nitrogen species modeled within the PRZM and VADOFT components of PRZM-3 are
presented. In Section 3, we explore On-site Wastewater Disposal Systems (OSWDSs). We
summarize the results of a computerized literature search, present the modeling strategy
that has been implemented in the OSWDS module, and describe database management
considerations related to modeling efforts utilizing both OSWDS and PRZM-3. Section 4
focuses on the capabilities that have been developed to evaluate the uncertainty associated
with modeling nitrogen processes within PRZM-3. Section 5 describes the initial model
testing efforts and conclusions and recommendations , and Section 6 provides literature
references. Appendices contain additional information pertaining to PRZM-3 input
description, sample PRZM-3 input, and OSWDS module input description. (
SECTION 2
SOIL NITROGEN PROCESS MODELING
2.1 SUMMARY OF LITERATURE REVIEW
A recent literature review (Donigian and Huber, 1991) identified a number of models
currently used for modeling nonpoint source loads from agricultural areas and concluded
that the HSPF AGCHEM and CREAMS/GLEAMS were the premier models for representing
soil nutrient processes in agricultural systems. Since that time, further application and
development work on both of these codes further confirms that these are the primary
candidates for potential inclusion into the PRZM-2 code for soil nitrogen modeling of septic
system discharges. Below we briefly discuss the current status of these models as related to
their capabilities for modeling soil nitrogen transformations and transport.
2.1.1 HSPF AGCHEM Nitrogen Modeling
The AGCHEM module of HSPF (Bicknell et al., 1993) attempts to represent the
major nitrogen and phosphorus transformation and transport processes occurring within the
soil profile that determine and control the fluxes of soil nutrients within the
soil/plant/terrestrial environment. This module maintains soil nutrient storages; allows for
various nutrient application modes; and simulates the nutrient balance and subsequent
movement of soil nutrients based on runoff, percolation, soil moisture, and sediment values
calculated by the corresponding sections of HSPF. These fluxes and state variables are used
by the AGCHEM module to provide the moisture storage and transport values needed for
the simulation of nutrient transformation and transport. Fertilizers, animal wastes, plant
residues, and other nutrient inputs (e.g., atmospheric deposition, sludge application) are
applied in their chemical form (i.e., NH4, NO3, PO4, Organic N, Organic P) either as a
-------
surface application or incorporated into the soil.
Nutrients are stored in four depth layers: surface zone, upper zone, lower zone, and
groundwater zone. The depth of each zone is defined by input parameters estimated by the
model user. The shallow surface layer is a continuous mixing zone that may be defined
functionally as the zone of interaction of surface-applied chemicals, from which surface
runoff, sediment erosion, and associated chemicals are transported to a waterbody. The
upper zone extends from the bottom of the surface zone to a depth often ranging from 10 to
20 cm, usually corresponding to the depth of major tillage or nutrient application
operations. The lower zone is the primary source of plant evaporation, and it regulates the
amount of soluble chemicals that can be introduced into groundwater since the chemicals
must pass through this layer. For agricultural applications, the lower zone depth is
typically set to the maximum depth of the crop root zone, usually ranging from 50 to 150 cm.
The groundwater layer represents the depth of shallow groundwater that actively
contributes baseflow to the stream channel. It can also be visualized as a shallow mixing
depth within the surface aquifer that controls the chemical transformations and associated
contributions to baseflow concentrations.
Within the AGCHEM module, the NITR section performs the reactions and
transformations of nitrogen species within the soil profile as a basis for predicting the soil
nitrogen storages and resulting nitrogen content of agricultural runoff, both surface and
subsurface. Under a joint US Geological Survey/EPA effort to improve AGCHEM for
representing N mass-balance modeling for forested areas at the watershed scale, a number
of refinements are being implemented and tested based on recommendations by Oak Ridge
National Laboratory as part of this joint effort (Hunsaker, et al., 1994). The version of
AGCHEM in HSPF Version No. 11 (Bicknell et a.,1 1995) includes these additional
algorithm refinements recommended for forested conditions (discussed below).
The nutrient simulation primarily assumes first-order reaction rates, but the recent
modifications have added additional options and capabilities to improve representation of
plant uptake, forest litter, and return of plant N to the soil. In the original version, the
processes simulated include immobilization, mineralization, nitrification/denitrification,
plant uptake, and adsorption/desorption (for which an equilibrium isotherm option is
available). The recent modifications include:
The particulate organic N state variable was divided into four state variables:
labile particulate, labile solution, refractory particulate, and refractory
solution. *
The plant N state variable was divided into aboveground and belowground
compartments.
A new pathway was added for plant N to "return" to organic N in the soil.
A litter N compartment was added to provide an intermediate compartment
between the aboveground plant N and the surface and upper layer organic N.
-------
A yield-based plant uptake option was included to provide an alternative to
the first-order rates and to allow for multiple cropping, nitrogen fixation, and
nutrient/moisture stress conditions.
Ammonia volatilization was added for conditions where animal waste or
fertilizer applications warrant this loss mechanism.
Reaction rates are input on a per day basis for each soil layer, which facilitates
compatibility with PRZM-2. Nitrite (N02) transforms so quickly in most agricultural soils
that it is not considered separately. The adsorbed phase represents the nutrients in a
complex form along with those adsorbed on the soil. Selected plant uptake parameters are
input monthly as a function of the stage of crop growth and expected yields. These monthly
parameters are adjusted to represent the crop uptake of N from the soil storages and to
distribute it throughout the growing season.
2.1.2 CREAMS/GLEAMS Models
Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS)
was developed by the U.S. Department of Agriculture's Agricultural Research Service
(Knisel, 1980; Leonard and Ferreira, 1985) for the analysis of agricultural best management
practices for pollution control. CREAMS is a field scale model that uses separate hydrology,
erosion, and chemistry submodels connected together using pass files.
Runoff volume, peak flow, infiltration, evapotranspiration, soil water content, and
percolation are computed on a daily basis. If detailed precipitation data are available,
infiltration is calculated at histogram breakpoints. Daily erosion and sediment yield,
including particle size distribution, are estimated at the edge of the field. Plant nutrients
and pesticides are simulated and storm load and average concentrations of sediment-
associated and dissolved chemicals are determined in the runoff, sediment, and percolation
through the root zone (Leonard and Knisel, 1984). User-defined management activities can
be simulated by CREAMS. These activities include aerial spraying (foliar or soil directed) or
soil incorporation of pesticides, animal waste management, and agricultural best
management practices (minimum tillage, terracing, etc.).
Groundwater Loading Effects of Agricultural Management Systems (GLEAMS) was
developed by the U. S. Department of Agriculture's Agriculture Research Service (Leonard
et al., 1987) to utilize the management-oriented, physically based CREAMS model and
incorporate a component for vertical flux "of pesticides. GLEAMS is the vadose zone
component of the CREAMS model.
GLEAMS, in its original version, consists of three major components, namely hydrol-
ogy, erosion/sediment yield, and pesticides. Precipitation is partitioned between surface
runoff and infiltration. Wwater balance computations are done on a daily basis. Surface
runoff is estimated using the modified Soil Conservation Service Curve Number Method.
The soil is divided into various layers, with a minimum of 3 and a maximum of 12 layers of
variable thickness used for water and pesticide routing.
10
-------
The most recent version of GLEAMS (Knisel et al., 1994) includes a greatly expanded
capability for modeling both soil nitrogen and phosphorus in agricultural systems. In this
new version, major processes in nutrient cycling and transformation were formulated and
incorporated into the code. GLEAMS now includes separate compartments for nitrate and
ammonia, along with four types of organic N -- active soil N, stable soil N, fresh organic N,
and animal waste organic N. The plant N compartment has been expanded to separate the
root, stover, and grain N components compartments are delineated for the surface layer (i.e.,
grain, stover, atmospheric N, and assimilated N), for both the surface and subsurface soil
layers (fresh organic N in crop residues and roots, fertilizer, nitrate, ammonia, and organic
N in animal waste), and for the active and stable soil N that occurs only in the soil. The
nitrogen processes include mineralization of the various organic N compartments,
nitrification, denitrification, ammonia sorption/desorption, nitrogen fixation, ammonia and
animal waste organic N volatilization, and plant uptake.
2.2 SOIL NITROGEN MODELING APPROACH IN PRZM-3
The HSPF/AGCHEM and GLEAMS nitrogen cycle modeling capabilities are similar
in many ways. Although the individual process representations differ, they both include
essentially the same processes and attempt to represent a complete nitrogen balance in each
soil layer and the entire soil profile. AGCHEM has probably been more widely applied since
the new GLEAMS version has been available for a limited time, but testing of both models
is still relatively limited. One apparent advantage of AGCHEM is the ability to handle
dissolved organic N (DON), which is of concern for septic system discharges; GLEAMS
appears to assume only particulate forms of all organic N, based on an initial review of the
documentation. DON has been added to AGCHEM in the most recent modifications
primarily to represent forested areas better, but the inability to model DON, and its
mobility, has been an identified limitation in earlier applications.
Because of this DON representation, the lack of any obvious advantage to GLEAMS
over AGCHEM, and our familiarity with the AGCHEM code, the AGCHEM NITR routines
were selected as the basic components for the nitrogen modeling capability in PRZM-3. We
did, however, review the appropriate sections of the GLEAMS model to identify any
algorithm refinements that could improve the nitrogen cycle representation. One of these
areas is the denitrification process; AGCHEM uses a first-order rate that can vary by soil
layer, whereas GLEAMS incorporates a soil moisture threshold so that denitrification is
only activated when soil moisture exceeds 110% of field capacity. A similar algorithm was
incorporated into PRZM-3.
i
2.2.1 Soil Nitrogen Storages and Transformations
Figure 2.1 shows the soil nitrogen storages and transformations included in PRZM-3.
As noted earlier, they are based on the soil nitrogen modeling procedures included in HSPF
AGCHEM Version No. 11, with a few additional changes to accommodate the PRZM soil
profile representation, a threshold for denitrification based on soil moisture, and the daily
timestep in PRZM. The nitrogen species of nitrate, ammonia, and four forms of organic
nitrogen [i.e. particulate organic nitrogen (labile and refractory) and dissolved organic
nitrogen (labile and refractory)] are represented. The soil nitrogen transformations include
11
-------
plant uptake of nitrate and ammonium, return of plant nitrogen to organic nitrogen,
denitrification or reduction of nitrate-nitrite, immobilization of nitrate-nitrite and
ammonium, mineralization of organic nitrogen, fixation of atmospheric nitrogen,
volatilization of ammonium, and the adsorption/desorption of ammonium and the organic
forms. All reactions and fluxes are computed on a daily basis and then the storages are
updated.
12
-------
Figure 2.1. PRZM-3 Soil Nitrogen Transformations
Uptake of N03 and NH4
Denitrification
Atmospheric
Deposition
Volatilization L
Atmospheric
Deposition
Sorption
Atmospheric
Deposition
Sorption
Sorption
* return of above ground plant and litter N occurs to
surface soil horizon only
13
-------
Nitrogen reactions can be divided between those that are chemical in nature and
those that are a combination of chemical and biological reactions. The adsorption and
desorption of ammonium is a chemical process. The user has the option of simulating
ammonium adsorption and desorption by first order kinetics with subroutine FIRORD or by
the Freundlich isotherm method with subroutine SV (discussed below).
The other reactions are a combination of biological and chemical transformations.
They all can be represented by first order kinetics, but plant uptake can optionally use
another algorithm (described below). The optimum first order kinetic rate parameter is
corrected for soil temperatures below 35 degrees C by the generalized equation:
KK = K*TH**(TMP-35.0)
where:
KK = temperature corrected first order transformation rate (per day)
K = optimum first order reaction rate parameter (per day)
TH = temperature coefficient for reaction rate correction (-)
(typically about 1.06)
TMP = soil layer temperature (degrees C)
Soil temperature must be simulated when nitrogen processes are being
simulated with PRZM-3. When temperatures are greater than 35 degrees C, the rate is
considered optimum, that is, KK is set equal to K. When the temperature of the soil layer is
below 4 degrees C or the layer is dry, no biochemical transformations occur. Identifiers with
a leading "K" (e.g., KDNI) are the optimum rates; those for corrected rates have both a
leading and current soil temperature value).
The biochemical reaction rate fluxes that are shown in Figure 2.1 are coupled, that
is, added to and subtracted from the storages simultaneously. The coupling of the fluxes is
efficient in use of computer time, but has a tendency to produce unrealistic negative
storages when large reaction intervals and large reaction rates are used jointly. A method
has been introduced that will modify the reaction fluxes so that they do not produce
negative storages. A warning message is issued when this modification occurs.
2.2.1.1
Ammonia Volatilization
The purpose for including ammonia volatilization is to allow large concentrations of
ammonia in the soil resulting from OSWDS inputs, animal waste, and fertilizer applications
to be attenuated by losses to the atmosphere. A simple, first-order rate is used with an
adjustment for air temperature. The original formulation, which was developed by Reddy et
al. (1979), was also adjusted for the soil cation exchange capacity (CEC) and wind speed,
and was automatically turned off after 7 days. In this implementation, we assume that the
constant CEC factor can be incorporated into the first-order rate constant by the user, and
that the wind (air flow) is always high enough to result in maximum loss (note: Reddy's
equation reduced volatilization only when wind was less than 1.4 km/day). Also, the user
14
-------
will be able to define volatilization rates for each soil horizon so that the rates can decrease
as the ammonia moves down through the soil column.
The volatilization flux in each layer is computed as:
AMVOL = AMSU * KVOL * TCVOL**(TEMP-20)
where:
AMVOL = loss of ammonia (mg/l/day)
AMSU = dissolved ammonia concentration (mg/1)
KVOL = rate constant at 20 C (/day)
TCVOL = temperature correction coefficient (-)
TEMP = air temperature (C)
The temperature correction for volatilization of ammonia is slightly different than
described above for the other first-order rate processes. The reference temperature is user-
specified, since rates in the literature are often given at a temperature of 20 degrees C (the
default). Also, the rate will be adjusted upwards when the soil temperature exceeds the
reference temperature.
2.2..1.2
Sorption/Desorption of Ammonium
When FORAFG = 0, the ammonium adsorption and desorption reaction fluxes of
chemicals are simulated with the FIRORD subroutine using temperature dependent first-
order kinetics. The calculation of the reaction fluxes by first-order kinetics for soil
temperatures less than 35 degrees C takes the form:
DBS = CMAD*KDS*THKDS**(TMP-35.0)
ADS = CMSU*KAD*THKAD**(TMP-35.0)
where:
DES = current desorption flux of chemical (mass/area per interval)
CMAD = storage of adsorbed chemical (mass/area)
KDS = first-order desorption rate parameter (per interval)
THKDS = temperature*correction parameter for desorption
TMP = soil layer temperature (degrees C)
ADS = current adsorption flux of chemical (mass/area per interval)
CMSU = storage of chemical in solution (mass/area)
KAD = first-order adsorption rate parameter (per interval)
THKAD = temperature correction parameter for adsorption
(THKDS and THKAD are typically about 1.06 )
All of the variables except the temperature coefficients may vary with the layer of
the soil being simulated. As noted above, soil temperature must be simulated when
15
-------
nitrogen is being simulated. The temperature correction of the reaction rate parameter is
based on the Arrhenius equation. At temperatures of 35 degrees C or above no correction is
made. When the temperature is at 0 degrees C or below or the soil layer is dry, no
adsorption and desorption occurs.
When FORAFG = 1, the SV subroutine simulates sorption/desorption based on the
Freundlich isotherm, which, unlike first-order kinetics, assumes instantaneous equilibrium.
That is, no matter how much chemical is added to a particular phase, equilibrium is
assumed to be established between the solution and adsorbed phase of the chemical. These
methods also assume that for any given amount of chemical in the soil, the equilibrium
distribution of the chemical between the soil solution and on the soil particle can be found
from an isotherm.
The adsorbed and solution phases of ammonium are determined in this subroutine
by a modification of the standard Freundlich equation (shown below). When the amount of
chemical is less than the capacity of the soil particle lattice to permanently bind the
chemical (XFIX), then all the material is considered fixed. All the fixed chemical is
contained in the adsorbed phase of the soil layer storage. Otherwise, the Freundlich
equation for curve 1 is used to determine the partitioning of the chemical into the adsorbed
and solution phases:
X = KF1*C**(1/N1) + XFIX (3)
where:
X = chemical adsorbed on soil (ppm of soil)
KF1 = single value Freundlich K coefficient
C = equilibrium chemical concentration in solution (ppm of solution)
Nl = single value Freundlich exponent
XFIX = chemical that is permanently fixed (ppm of soil)
The above equation is solved in subroutine ITER by an iteration technique. The parameters
used in the computation can differ for each soil horizon.
2.2.2 Nitrogen Inputs: OSWDS Loadings, Atmospheric Deposition, Nitrogen Applications
Inputs of nitrogen to the surface and subsurface soil horizons can be accommodated
for representing OSWDS loadings, atmospheric deposition, and nitrogen additions through
fertilizer and/or manure applications. All nitrogen inputs are defined in their elemental
forms as N03-N, NH4-N, and organic N; for each of the three input categories, further
restrictions may apply on the form and species of the applied amounts (discussed below).
OSWDS loadings can be provided as user-defined WDM files, or as output files
generated by either or both of the treatment options included in the OSWDS module; they
are then input to a specific PRZM soil horizon defined by the user.
Two basic types of atmospheric deposition are simulated. Dry deposition is
considered to be a flux per unit area over the land surface that is independent of rainfall.
16
-------
Wet deposition is considered to be a concentration of a nitrogen species dissolved in the
input precipitation. If data are available as a total flux only, they should be input as dry
deposition. All deposition inputs are added to the surface soil horizon and are assumed to
be input as NO3-N, adsorbed NH4-N, and particulate labile organic N.
If atmospheric deposition is being simulated, the soil storage in the surface horizon is
updated for each of these three species of nitrogen using the formula:
N(i+l) = N(i) + ADFX + PREC*ADCN (4)
where:
N(i) = storage of nitrogen species in the soil layer on day i, in
mass/area
ADFX = dry or total atmospheric deposition flux in mass/area per
interval
PREC = precipitation depth
ADCN = concentration for wet atmospheric deposition in mass/volume
Nitrogen applications with fertilizers or manure is accomplished ii\ a manner
analogous to pesticide applications. Application dates are specified for the entire simulation
period, along with the specific amounts of each N form ~ NO3-N, NH4-N, and organic N ,
the depth of incorporation for each application, and the labile fraction of the applied organic
N.
2.2.3 Plant Uptake
Plant uptake of soil nutrients in PRZM-3 can be modeled by two alternative methods
using the NITR module. When NUPTFG = 0, plant uptake is represented the same as in
the AGCHEM module in HSPF Version No. 10, as a first-order rate process with an
Arrhenius temperature correction adjustment based on simulated soil temperatures. The
first-order plant uptake rates are defined by the user, can be specified separately for each
soil horizon within PRZM, and can vary for each month to approximate the monthly pattern
of crop growth and nutrient uptake. The rates are adjusted during calibration to mimic the
expected annual nutrient uptake and the seasonal pattern for the specific crop and practice.
Plant uptake can be distributed between nitrate and ammonium by input parameters
intended to designate the fraction of plant uptake from each species.
Because this option uses first-order monthly uptake rates to represent time-varying
plant nutrient uptake, the calculated uptake amounts are highly sensitive to, and a direct
function of, the available nutrients in the soil profile and the specific nutrient
input/application rates. This causes a problem when application rates are changed, such as
under nutrient reduction alternatives, because the uptake amounts are not a function of
expected crop yields and associated nutrient uptake; thus, even though sufficient nutrients
may be available to satisfy crop needs under the reduced application rates, the calculated
uptake may be less than the crop needs because of the first-order formulation.
As part of a previous effort, we reviewed the problem and issues related to the plant
uptake algorithms in the AGCHEM modules of HSPF, along with the primary alternative
17
-------
algorithms used in a number of current agricultural nutrient models (Donigian et al., 1995).
Based on that review effort and the compatibility of alternative functions with the
AGCHEM and HSPF soil profile representation, we selected the conceptual approach of the
plant uptake formulation in the NLEAP model (Shaffer, Halvorson et al., 1991) for
incorporation into AGCHEM/HSPF Version No. 11 (Bicknell et al, 1995). This selection was
based on the following characteristics of the NLEAP plant nutrient uptake function:
Calculates crop nutrient needs as a function of expected crop yield
Allows seasonal uptake variation based on expected crop growth patterns
Accommodates (or can be modified to accommodate) timesteps less than 1 day
Considers both NO3-N and NH4-N as available for N uptake
Considers N fixation, double cropping, and uptake from different soil layers
Allows the N uptake functions (except for N fixation) to be adapted for P
uptake in AGCHEM
Provides an overall level of detail that is consistent and compatible with
AGCHEM
Due to differing hydrology, soil moisture, and soil profile simulation procedures
among NLEAP, AGCHEM, and PRZM, the NLEAP plant N uptake functions required
adaptation in order to be consistent and compatible. The changes primarily provided
greater user flexibility in defining the timing and distribution of plant uptake from the
individual soil layers, whose depths are user-specified in both AGCHEM and PRZM, and to
represent a wider potential range of land surface conditions. The details of the changes are
discussed by Donigian et al. (1995).
The yield-based plant nitrogen uptake formulation is selected when NUPTFG = 1,
and is essentially the same in PRZM-3 and AGCHEM/HSPF Version No. 11. A total
annual target, NUPTGT, is specified by the user, and is then divided into monthly targets
during the crop growing season for each soil horizon. The monthly target for each horizon is
calculated as:
MONTGT = NUPTGT* NUPTFM(MON)* NUPTM(MON)* (5)
CRPFRC(MON.ICROP)
where:
MONTGT = monthly plant uptake target for current crop, mass N/area
NUPTGT = total annual uptake target, mass N/area
NUPTFM = monthly fraction of total annual uptake target, dimensionless
NUPTM = soil horizon fraction of monthly uptake target, dimensionless
CRPFRC = fraction of monthly uptake target for current crop,
dimensionless. This is 1.0, unless the month contains parts of
two or more crop seasons, in which case the monthly uptake
18
-------
target is divided among the crops according to the number of
days of the month belonging to each crop season.
MON = current month
ICROP = index for current crop
Planting and harvesting dates can be specified for up to three separate crops during
the year. Plant uptake is assumed to occur only during a growing season, defined as the
time period between planting and harvest. When portions of two growing seasons are
contained within 1 month, the total monthly target is divided between the two crops in
proportion to the number of days in each season in that month. The daily target is
calculated by starting at zero at the beginning of a crop season and using a trapezoidal rule
to solve for monthly boundaries; linear interpolation is used to solve for daily values
between the monthly boundaries, and between a monthly boundary and a planting or
harvest date.
Yield-based plant uptake only occurs when the soil moisture is above the wilting
point, which is specified by the user for each soil horizon, and sufficient nutrients are
available. No temperature rate adjustment is performed, but all uptake is stopped when
soil temperature is below 4 degrees C. If the uptake target is not met during a given
interval, whether from nutrient, temperature, or moisture stress, then a deficit is
accumulated, and applied to the next interval's target. When uptake later becomes possible,
the program will attempt to make up the deficit by taking up nitrogen at a rate higher than
the normal daily target, up to a user-specified maximum defined as a multiple of the target
rate. The deficit is tracked for each soil layer, and is reset to zero at harvest; i.e., it does
not carry over from one crop season to the next.
When using the yield-based plant uptake option, it is also possible to represent
leguminous plants (e.g. soybeans) that will fix nitrogen from the atmosphere. The
algorithm is designed to allow N fixation only to make up any shortfall in soil nitrogen; i.e.,
fixation is only allowed if the available soil nitrogen (i.e., nitrate and solution ammonium) is
insufficient to satisfy the target uptake. The maximum daily nitrogen fixation rate is
subject to the same limits as the uptake under deficit conditions noted above.
2.2.4 Soil and Plant Nitrogen, and Litter Compartments
In the previous version of the NITR module in HSPF AGCHEM, plant N was a single
"state variable" that represented the cumulative amount of N taken up by plants from each
soil layer. This material continues to "build up" during the simulation; i.e., it is not
converted to any other species in the soil. This option has been incorporated into PRZM-3.
In addition, we added a pathway in each layer so that plant N can be converted (by first-
order rate) to organic N (labile particulate) to represent the return of plant N to the soil
through leaf fall or crop residues. This rate can be either constant or monthly variable.
Also, nitrogen that is taken up by the plants can be divided between above ground
and below ground fractions (using a simple fraction of the total uptake). Above ground plant
N returns to the litter compartment, and litter N returns to particulate organic N (with
labile and refractory fractions) in the surface soil horizon. Both of these reactions are
19
-------
simulated using first-order kinetics. No other reactions affect these nitrogen storages except
for plant uptake to the above ground compartment, as calculated in subroutine NITRXN.
Return of litter and below ground plant N to particulate organic N is divided into
labile and refractory fractions, which can be constant or monthly variable. Regardless of the
option used to simulate plant uptake, if the above ground and litter compartments are being
simulated, then the user can specify the fraction of uptake from each layer that goes to the
above ground storage. The rest is assumed to remain within the below ground plant N
compartment for that soil layer.
2.2.5 Organic Nitrogen Compartments and Reactions
The previous NITR module of AGCHEM contained a single organic N state variable
in each soil layer. This material was assumed to be a particulate species that is increased
from immobilization of nitrate and ammonia, and is converted back to ammonia by
mineralization in the soil. It also is transported on the soil surface by association with
sediment. In the new module, this species is described as a "particulate labile" fraction of
organic N; it will undergo conversion by first order rate to a "particulate refractory" fraction,
and it will partition to a "soluble labile" fraction. The "particulate refractory" species will
also partition to a "soluble refractory" fraction. The two soluble species, therefore, will be
available for transport as runoff and leaching within the soil profile, and likewise, the new
particulate fraction will be transported on the surface with the eroded soil. The partitioning
reactions are described by a simple ratio of particulate concentration to solution
concentration, i.e., a standard linear partition coefficient. The four fractions and their
assorted reactions are illustrated in Figure 2.1. Note that the storages and transformations
in this figure are repeated in each soil horizon except for the aboveground plant N and the
litter compartments.
2.3 MODELING APPROACH WITHIN VADOFT COMPONENT
While a detailed representation of nitrogen species and processes is warranted
within the upper subsurface area modeled using the PRZM component, it is appropriate to
implement a more limited subset of nitrogen processes to represent chemical behavior
within the lower VADOFT zone of modeling. As the OSWDS and PRZM-3 testing effort in
Section 7 describes, a typical model application would make use of the PRZM component of
PRZM-3 to represent the top 7 to 9 feet of the soil column, with the VADOFT component
used to model depths below 7 to 10 feet. At these lower depths, processes such as
nitrification, volatilization and ammonification are much less prevalent than in areas closer
to the surface.
Therefore, our modeling approach has been to limit detailed nitrogen algorithm
enhancements to the PRZM component, and then enable the individual species fluxes to be
advected through the VADOFT portion, allowing the user to represent 'decay/loss' and
'retardation' in a manner that parallels pesticide modeling using VADOFT. The VADOFT
decay parameter provides a means of representing denitrification of nitrate in the event
that saturated soil conditions associated with clay materials prevail within the VADOFT
soil zone. (Note, however, that the use of the VADOFT decay parameter will result in a
20
-------
constant loss of chemical that is not tied to the transient occurrence of saturated conditions.)
The VADOFT retardation coefficient provides a means, via a less mechanistic
method than that used in PRZM, to represent the effects of adsorption of nitrogen species to
soil in the VADOFT compartment. In particular, retarding the movement of ammonium
relative to flow may be warranted. If the user opts to use VADOFT, ammonia/ammonium,
nitrate/nitrite, and total organic N are all automatically simulated; consequently the
VADOFT input sequence developed using the input description on pages 4-25 through 4-38
of the PRZM-2 manual must provide parameters for all three of these nitrogen pools.
SECTION 3
ON-SITE WASTEWATER DISPOSAL SYSTEM
NITROGEN EFFLUENT MODELING
3.1 SUMMARY OF LITERATURE REVIEW
3.1.1 General Description of On-site Wastewater Disposal Systems(OSWDS)
The vast majority of on-site wastewater disposal systems (OSWDSs) in the United
States are conventional gravity flow systems (Reneau et al., 1989). The components of the
system are a septic tank, a distribution box, and a network of drainfield trenches. The
system normally relies on gravity (1) to carry household waste to the septic tank, (2) to move
effluent from the septic tank to the distribution box, and (3) to distribute effluent from the
distribution box throughout the drainfield. The septic tank provides primary treatment for
wastewater; as such, it acts as a settling chamber, provides storage for sludge and scum,
and reduces the quantities of organic materials transported to the distribution, treatment,
and disposal components of the system. Anaerobic digestion in the septic tank results in a
reduction of sludge volume (40%), biological oxygen demand (60%), suspended solids (70%),
and conversion of much of the organic N to the NH4+ form (75-85%).
3.1.2 Residential Septic Tank Influent Characteristics
Septic tanks are used primarily for treatment of wastes from individual residences.
In rural areas, they are also used for establishments such as schools, summer camps, parks,
trailer parks, and motels (Metcalf & Eddy, Inc., 1972). Given this breadth of use,
characterization of the flow of wastewater into a septic tank depends on the number of users
and an estimate of per capita loading per day. EPA (1980) has compiled extensive data
characterizing ranges and median values for wastewater volume and nitrogen species
concentrations. Eight flow studies involving 109 residences resulted in a weighted average
loading of 44 gallons/capita/day, with the range of study averages being 36 to 52
gallons/capita/day. These values constitute the sum of all wastewater, black and gray, that
is introduced into septic tanks. (Blackwater is defined as liquid and solid human body
waste and the carriage waters generated through toilet usage. Graywater is defined as
wastewater generated by water-using fixtures and appliances, excluding the toilet and
21
-------
possibly the garbage disposal.)
With regard to nitrogen loadings in the wastewater influx to residential septic tanks,
EPA (1980) estimated mean values of 140 mg/1 total N for toilets, 79 mg/1 for garbage
disposals, and 17 mg/1 for basins, sinks, and appliances. With the typical usage patterns,
EPA estimates a combined wastewater mean concentration of total N of 63 mg N/l. (Note
that comparison of this value with estimates of septic tank effluent N characteristics
available in the literature [i.e., 72 mg N/l (NVPDC, 1990); 40-80 mg/1 (Reneau et al., 1989)]
suggests that septic tank processes result in minimal loss of total N from the system.
Rather, the function of the tank is temporary storage and transformation. EPA has
determined that the breakdown of influent N consists of approximately 82% organic N, 18%
ammonia/ammonium, and negligible nitrite/nitrate. Minnesota State EPA has published
similar estimates for N distribution of 78% organic N, 19% ammonia/ammonium, and 1%
nitrite/nitrate (Wall, 1991). The reported range of influent concentration values are 29-82
mg/1 organic N, 6-18 mg/1 NH4+, and <1 mg/1 NCy.
3.1.3 Nonresidential Septic Tank Influent Characteristics
*
EPA (1980) has compiled and/or developed estimates of typical wastewater flow for
13 types of commercial sources, 8 types of institutional sources, and 16 types of recreational
sources. With regard to N loadings to nonresidential septic tanks, EPA believes that there
is simply too much variability to develop meaningful estimates for the various source types.
In many vacation-oriented areas, occupancy rates, and hence loadings to septic tanks
(residential and nonresidential), are highly seasonal in nature. In such areas, proper
estimation of loadings requires consideration of seasonal occupancy.
3.1.4 Septic Tank Effluent Characteristics
For a properly functioning septic tank, inflows should very closely match outflows
(i.e., loss of flows due to leakage, evaporation and/or pumping is insignificant). Values for N
species concentrations in the effluent from conventional domestic septic tanks fall within a
reasonably well-defined range. Estimates identified in the literature are provided in Table
3.1.
Table 3.1. Typical Mean Concentration Values (mg/1) for Nitrogen Species in Septic Tank
Effluent
EPA (1980)
Reneau (1989)
NVPDC (1990)
Total N
NH4+-N
NCy-N
Dissolved Org. N
Particulate Org. N
46
40-80
30-60
10-20
72
60
12
22
-------
3.1.5 Septic Tank Pumping
The Septic Systems Handbook (Kaplan, 1991) estimates that approximately 4% of
the total N entering a septic tank is removed via pumping, if once-every-3-years pumping
scheduleis assumed. This estimate is consistent with the fact that the total N
concentrations of septic tank influent and effluent appear very similar. Minnesota State
EPA has published a high-end estimate of pumping removal efficiency as 10% (Wall, 1991).
3.1.6 Septic Tank Failure
Detailed descriptions of the cause and nature of septic tank failures are available in
reports prepared by the Northern Virginia Planning District Commission (1990,1992). It is
important to remember that failure connotes either backup of wastewater into the residence
or "daylighting" of wastewater on the surface above the tank and leachfield due to drainfield
saturation. The important conclusion of these reports is that the relative amount of effluent
that is released to the soil surface as a result of failed conditions is very small (0.2 to 0.3 %
of total effluent), and hence failure can probably be ignored in modeling exercises focused on
evaluating potential for groundwater contamination. (Note: The data evaluated for the
NVPDC studies included all septic tanks in the region including a substantial number that
were more than 30 years old, and tanks located in soils with a wide variety of percolation
characteristics; there is no immediate reason to believe that the occurrence of failure is
significantly greater in other geographic areas.) Ignoring the effects of septic tank failures
in a N modeling scheme is also supported by the fact that the vast majority of the effluent
that is actually released to the soil column during a condition considered "failure" has
already spent its residence time in the tank, performed the traditional transformations and
entered soil column in much the same way that it would normally; however, after the
effluent enters the soil, it is unable to percolate and instead "ponds" until it reaches the soil
surface. It is reasonable to assume that from the point of view of estimating N species
loadings to the soil column, such a failure does not have a significant effect.
3.1.7 Septic Tank Models and Modeling
The modeling of septic tanks is not a common practice. The lack of septic tank
models and modeling activities can most likely be attributed to the relative simplicity,
consistent functioning, and isolation of the devices. As long as the tank is cleaned once
every 3-5 years and is not overloaded with influent (or subjected to non-typical influents),
studies suggest that it effectively removes scum and solids, and transforms approximately
three-fourths of the influent organic N into ammonium. The literature consistently
suggests that variability in the contribution of N species to groundwater depends mostly on
(1) design options and conditions within the distribution component of the OSWDS and (2)
the hydrogeological setting (i.e., soil stratigraphy, macropore flow opportunities, depth to
water table) of the tank. Thus, while the septic tank has remained essentially unchanged in
design for decades, current planning efforts to manage nitrate loadings to groundwater
focus on issues of allowable density of tanks, and on optimal siting of individual tanks; and
current engineering efforts to mitigate nitrate contamination of groundwater focus on
modifications of the traditional distribution/leachfield mechanism.
3.1.8 Septic Tank Effluent Distribution/Treatment Systems
23
-------
The conventional distribution system for septic tank effluent consists of a
distribution box, which separates flow evenly into a network of drainfield trenches.
Drainage holes at the bottom of each line allow the wastewater to drain into gravel trenches
for temporary storage. (In some areas "seepage pits" have also been in common use as an
alternative to a leachfield; these devices provide a less-distributed temporary storage of
septic tank effluent, with gradual percolation into the soil.) In reality, uneven distribution
of effluent in conventional OSWDSs results in elevated loading rates to a relatively small
portion of the individual trench, and to a limited number of trenches. These elevated
loading rates increase the potential for contamination to groundwater because of saturated
flow conditions and can, with time, decrease infiltration because of accelerated formation of
a biological clogging mat (Bouma et al., 1972) and accumulation, via filtration, of the limited
amount of solids carried by effluent (DeVries, 1972). In slowly permeable soils the clogging
mat may result in inadequate infiltration and ponding. Efforts to mitigate this problem,
known as creeping failure, have led to use of several new mechanisms including low
pressure distribution systems (Reneau et al., 1989) and distribution box flow diversion
valves (NVPDC, 1992).
Additional alterations of the conventional distribution system are directed at
encouraging the treatment of wastewater via the coupled processes of
nitrification/denitrification. To do so, the homogeneous gravel trenches that traditionally
surround drainfield trenches are being replaced by numerous combinations of materials
that encourage nitrification of ammonium, followed by denitrification of nitrate-nitrite.
Other schemes have been utilized to arrest nitrogen via methods such as metal
complexation. Schemes that utilize such mechanisms as reactive porous media barriers
(Robertson and Cherry, 1995), sphagnum peat (Brooks et al., 1984), multi-soil layering with
iron and pelletized jute (Wakatsuki, 1993) are fairly common in the literature. Another
alternative system of note is the use of low pressure distribution in conjunction with
surficial mounds to increase the distance between the treatment area and shallow water
tables.
(Note: The primary relevance of all these alternative distribution/treatment systems is that
significant chemical transformations and/or physical arrest of nitrogen can occur after the
wastewater leaves the septic tank and before the wastewater enters the undisturbed
subsurface soil. Hence, there is a need to represent these processes (i.e., transformation and
removal) a second time within the OSWDS module prior to introducing the effluent into a
PRZM-3 compartment to model traditional soil processes.)
3.1.9 OSWDS - Soil Profile Interface
As the distribution/treatment system component of OSWDSs become more highly
engineered, identifying the interface between the OSWDS and the soil becomes more
difficult. For example, the above-mentioned "reactive porous media barrier (actually, a
layer of sawdust)" is installed in the vadose zone below the infiltration bed of an otherwise
conventional system. For a mounded system, the interface is likely the ground surface.
3.2 MODELING APPROACH
The appropriate level of detail for representing OSWDS inflows/processes/outflows is
achieved by implementing a generalized module (i.e., one not dependent on specific reaction
kinetics) comprised of two treatment units (Figure 3.1). The first treatment unit always
24
-------
represents the septic tank; the second treatment unit represents all transformations/ losses
that occur between the outlet of the septic tank and the inflow into the unaltered subsurface
soil. (For our purposes, "unaltered" means below or beyond the area that has been modified
for purposes of wastewater distribution and/or treatment.)
25
-------
Figure 3.1. OSWDS Schematic
Nitrogen Transformation Factors for "Typical" Scenario Test Run
Q
M ORGN
M NH4
M N03-N02
1st Treatment Unit Q
M ORGN
M NH4
M N03-N02
2nd Treatment Unit
Q
M ORGN
M NH4
M N03-N02
Transformations/Losses
Transformations/Losses
ORGN
NH4
NO3-N02
ORGN
.296
0.00
0.00
NH4
.054
1.0
0.00
NO3-NO2
0
0
1
LOSS1
.056
0.00
0.00
ORGN
NH4
N03-NO2
ORGN
1.0
0.0
0.
NH4
0.0
1.0
0.00
NO3-NO2
0.0
0.0
1.0
LOSS1
0.00
0.00
0.00
LOSS2
0.0
0.0
0.0
LOSS3
0.0
0.0
0.0
LOSS1 = Settling, Physical Removal
LOSS2 = Denitrification
LOSS3 = Volatilization
26
-------
The user has the option of whether to consider only the first treatment unit, or both
treatment units. If only one treatment unit is modeled, the output flow and N
concentrations from the unit are directly input into the appropriate PRZM-3 soil
compartment. The PRZM-3 soil horizon into which OSWDS outflow is introduced is
specified by the user based on knowledge of the effluent depth compared to the soil horizon
depths. If only one treatment unit is modeled, this depth corresponds to the depth below the
ground surface of the tank outlet; if two treatment units are modeled, the effluent depth
typically corresponds to the bottom depth of the area modified for wastewater
distribution/treatment. Effluent is assumed to be homogeneously distributed throughout the
horizon into which it is introduced.
Wastewater influent to the first treatment unit is characterized according to the following
scheme:
(1) The user defines a "base" timeseries of wastewater flow (gal/day) and
concentrations of associated nitrogen species (mg/1). The base wastewater
flow is defined by assigning a value for per person wastewater generation
(gal/capita), and one or more seasonal occupancy rates (# of persons serviced
by the OSWDS).
(2) The ability to define seasonal occupancy is enabled. The user specifies the
number of "occupancy seasons" during the calendar year, the starting date of
each season, and the number of occupants serviced by the OSWDS.
(3) The modeling scheme assumes that the nitrogen species concentrations
associated with a particular OSWDS remain constant over time (i.e., the flows
can vary seasonally, but the concentrations do not vary with time.) The user
defines the concentrations of N species in the influent; reasonable default
values gleaned from the literature for residential systems are provided in
Table 3.2.
Table 3.2. Typical Mean Concentration Values (mg/1) for Nitrogen Species in Septic Tank
Effluent
EPA (1980)
Reneau(1989)
NVPDC (1990)
Total N
NH/-N
NO3'-N
Dissolved Org. N
Particulate Org. N
46
40-80
30-60
10-20
72
60
12
(4) Nitrogen species that are modeled in the influent wastewater are total
organic N, ammonium, and nitrate-nitrite. While the literature consistently
reports negligible amounts of nitrate-nitrite in typical wastewater influent,
27
-------
for the sake of flexibility combined nitrate-nitrite is included as a possible
wastewater constituent for atypical situations. Given that the modeling of N
transformations within the OSWDS module is not mechanistic, there is no
benefit to differentiating between particulate and dissolved organic N, or
between labile and refractory organic N. This distinction, however, is needed
prior to input to the soil region represented by PRZM-3 (see discussion below).
The treatment effects of the septic tank are modeled as follows:
(5) The efficiency of the septic tank is defined by the user by (i) assigning values
for a series of transformation factors between N species and («) defining a
physical loss term for organic N due to settling/storage within the tank and
tank maintenance activities (i.e., pumping).
(6) N species that are modeled within the septic tank are total organic-N,
ammonium, and nitrate-nitrite. (Nitrate-nitrite concentrations are
consistently reported at insignificant levels within septic tanks, but
nonetheless, for the sake of generality, we will include transformation factors
that allow user-controlled specification of this constituent.
The treatment effects of the second treatment unit are modeled as follows:
(7) As in the first treatment unit, the efficiency of the second unit is defined by
the user by (i) assigning values for a series of transformation factors between
N species and (u) defining a physical loss term for organic N. The
transformation factors are expanded to allow representation of the
production, and loss of, elemental nitrogen via denitrification, and ammonia
via volatilization. (Literature suggests that transformation of ammonium to
nitrate-nitrite can be significant within the distribution/treatment area
outside the septic tank, particularly in systems engineered to facilitate
nitrification/denitrification.) The physical loss term represents the sum of
loss due to settling, clogging, complexation or any other process that results in
permanent physical arrest of nitrogen species within the confines of the
distribution/treatment area.
(8) N species that are modeled as state variables within the second treatment
unit are total organic-N, ammonium, and nitrate-nitrite.
Regarding the N constituent linkage between the OSWDS module and the PRZM-3
soil compartment, we have borrowed the modeling strategy used for modeling sediment in
the HSPF model (Bicknell et al., 1993). In the same manner that sediment is modeled as a
single constituent in the HSPF land surface module (PERLND) and then divided into sand,
silt and clay fractions (via user input) prior to its input in the HSPF instream module
(RCHRES), total organic N is modeled as a single constituent throughout the OSWDS
module. Capabilities are implemented for user-defined allocation of total N into particulate
labile and particulate refractory to parallel the N species scheme that is used in the PRZM-2
soil compartments. (All organic N effluent is assumed to be particulate.)
The modeling approach assumes that all OSWDSs are located in the subsurface area
that is represented by the PRZM component of PRZM-3 (i.e., septic tanks do not generate
28
-------
direct fluxes to VADOFT). The linkage has required the development of capabilities for
representing lateral influxes of both water and chemical constituents into specific PRZM-3
compartments. The soil horizon into which the lateral flows occur is user-defined.
Specification of effluent flow into a soil layer that is below the area modeled using the
PRZM component (i.e., in the area modeled using VADOFT) is not allowed, and results in an
error message and termination of the run.
3.3 DATABASE MANAGEMENT
The OSWDS module has been implemented with the dual capability to (1) write to
user-defined files, or (2) to interact with the ANNIE/WDM capabilities for timeseries
management and display of relevant flow and nitrogen species. Users are able to provide
the timeseries influent to the first treatment unit of the OSWDS module by defining flows
and concentrations in the module input sequence. Users are able to provide the timeseries
influent to the second treatment unit of the OSWDS module by direct use of values
generated by the simulation of the first treatment unit. The design assumes that all
interactions between the OSWDS module and the PRZM-3 model occur via PRZM-3 reading
OSWDS module output files to obtain input to the soil compartment(s). These files contain
flow/chemical mass flux data derived from one of two different run options: (1) results
generated by a simulation that only considers the first treatment unit and (2) results
generated by simulating both treatment units. Running the OSWDS module "stand-alone"
allows the user to develop scenarios related to different septic tank influent conditions (e.g.,
occupancy rates, seasonal occupancy) and/or treatment options (e.g.,
nitrification/denitrification schemes); store the results of the scenarios; and use the results
as input to various PRZM-3 soil horizon conditions.
SECTION 4
UNCERTAINTY ANALYSIS
PRZM-2 can be run in a Monte Carlo mode so that probabilistic estimates of
pesticide loadings to groundwater from a source area can be made. This same capability
has been added for the nitrogen species being modeled in PRZM-3.
The input preprocessor allows the user to select distributions for key parameters
from a variety of distributions the Johnson family (which includes the normal and
lognormal), uniform, exponential and empirical. If the user selects distributions from the
Johnson family, he or she may also specify correlations between the input parameters.
Section 8 of the PRZM-2 User's Manual (Mullins et al., 1993) devotes 18 pages to explaining
the options available for evaluating uncertainty; this information is equally applicable and
useful for users of PRZM-3. Note that the following section is devoted to discussion of
uncertainty analysis related directly to soil nitrogen input and output variables. It is quite
possible that a user may be interested in evaluating the effect of additional input variables
(e.g., soils characteristics) on output nitrogen fluxes. Such analyses can be performed using
PRZM-3, but the user will need to refer to the PRZM-2 manual (in particular pages 4-39
through 4-44) to obtain the necessary information to develop the PRZM-3 input sequence.
4.1 IMPLEMENTATION
29
-------
The Monte Carlo analysis process within PRZM-2 is well designed, making the
addition of new variables available for uncertainty analysis a straight-forward process. A
separate input sequence defines which input variables are to be varied, which output
variables are to be observed for statistical analysis, and number of simulations that are to
be performed. Additional parameters in this file define the distribution to be used for input
variables and the type of output to generate for the output variables.
As is the case with PRZM-2, PRZM-3 reads the standard input sequences (as for a
deterministic run) and replaces the appropriate input variables with Monte Carlo generated
values for each of the specified number of runs. Code has been added to use the new soil
nitrogen module's input variables in the existing Monte Carlo input modules. This has been
done in the same manner as for the existing Monte Carlo input variables.
The output processor of the Monte Carlo code prepares statistics for the output
variables specified in the Monte Carlo input sequence. These output statistics include mean
and maximum values, quantiles of the output distribution, and tabulation of cumulative
frequency histograms. Code has been added to transfer the new nitrogen module's outputs
to the arrays used by the Monte Carlo output modules. This has been done in the same
manner as for the existing Monte Carlo output variables.
4.2 INPUT VARIABLES
A linkage of the OSWDS module to the existing PRZM-2 Monte Carlo capabilities
has not been performed. Hence, capabilities to perform Monte Carlo analysis are restricted
to nitrogen parameters contained within PRZM-2. A number of input variables in the soil
nitrogen module have been included as uncertainty analysis inputs. Table 4.1 provides a
list of these variables and the Monte Carlo labels used to specify each variable in the Monte
Carlo input sequence. Monte Carlo input description is provided on pages 4-40 through 4-42
of the PRZM-2 manual.
Table 4.1. Monte Carlo Nitrogen Variable Input Options for PRZM-3
Parameter Monte Carlo Label
Nitrate Application (kg/ha) N03 APPLICATION
Ammonia Application (kg/ha) NH3 APPLICATION
Organic N Application (kg/ha) , ORGN APPLICATION
Plant N Uptake Rate (/day) PLANTN UPTAKE
Below-Ground Plant N Return Rate (/day) BG PLANT N RETURN
Above-Ground Plant N Return Rate (/day) AG PLANT N RETURN
Ammonium Desorption Rate (/day) NH4 DESORPTION
Ammonium Adsorption Rate (/day) NH4 ADSORPTION
Nitrate Immobilization Rate (/day) N03 IMMOBILIZATION
Organic N Ammonification Rate (/day) AMMONIFICATION
Denitrification Rate (/day) DENITRIFICATION
Nitrification Rate (/day) NITRIFICATION
Ammonium Immobilization Rate (/day) NH4 IMMOBILIZATION
Ammonia Volatilization Rate (/day) NH3 VOLATILIZATION
30
-------
Refer to page 4-39 of the PRZM-2 manual for an example input sequence, and to
pages 4-40 through 4-42 for instructions on developing a new input sequence. The
"parameter," "Monte Carlo label" and "index" information for additional input parameters
(e.g., soil bulk density, wilting point, field capacity) that are contained in on page 4-43 of the
PRZM-2 manual can be used directly in developing Monte Carlo input sequences for PRZM-
3.
Note that the new Monte Carlo nitrogen variable input options have been limited to
the PRZM component of PRZM-3; there are no new options for performing Monte Carlo
analysis of nitrogen input variables within the VADOFT component. PRZM-3, however,
does allow the user to perform Monte Carlo analysis on certain soil-related input variables
(see page 4-44 of PRZM-2 manual), and to determine the effect of that uncertainty on
nitrogen species fluxes.
4.3 OUTPUT VARIABLES
The ultimate objective of the new nitrogen modeling capabilities is fo project
nitrogen loadings to groundwater. Thus, the primary output variables that need to be
available for Monte Carlo output analysis are loadings of the different species at the bottom
of the PRZM-3 column. For pesticides, this is typically reported as a core flux, in g/cm2/day,
from the bottom compartment of PRZM-3 (either at the bottom of PRZM or VADOFT,
depending on whether a VADOFT component is simulated or not). In the same manner,
core flux values for the different nitrogen species are now included as Monte Carlo output
variables. The output options for nitrogen species fluxes, and for nitrogen concentrations at
the bottom of VADOFT, are detailed in Table 4.2 below.
31
-------
Table 4.2. Monte Carlo Nitrogen Variable Output Options for PRZM-3
Parameter
Monte Carlo Label
Random PRZM Model Outputs
Runoff Flux, Ammonia
Runoff Flux, Nitrate
Runoff Flux, Organic N
Erosion Flux, Ammonia
Erosion Flux, Organic N
Groundwater Flux, Ammonia
Groundwater Flux, Nitrate
Groundwater Flux, Organic N
Groundwater Flux, Total N
Plant Uptake Flux, Ammonia
Plant Uptake Flux, Nitrate
Plant Uptake Flux, (NH3 + NO3)
Plant Return Flux, Organic N
Immobilization Flux, Ammonium
Immobilization Flux, Nitrate
Immobilization Flux, (NH4 +N03)
Volatilization Flux, Ammonia
Denitrification Flux
Nitrification Flux
Ammonification Flux
Random VADOFT Model Outputs
Advection Flux, Ammonia
Advection Flux, Nitrate
Advection Flux, Total Organic N
Dispersion Flux, Ammonia
Dispersion Flux, Nitrate
Dispersion Flux, Total Organic N
Decay Flux, Ammonia
Decay Flux, Nitrate
Decay Flux, Total Organic N
Concentration, Ammonia
Concentration, Nitrate
Concentration, Total Organic N
RUNOFF FLUX NH3
RUNOFF FLUX NO3
RUNOFF FLUX ORGN
EROSION FLUX NH3
EROSION FLUX ORGN
GW FLUX NH3
GW FLUX NO3
GW FLUX ORGN
GW FLUX TOTN
UPTAKE FLUX NH3
UPTAKE FLUX NO3
UPTAKE FLUX TOTN
RETURN FLUX ORGN
IMMOBIL. FLUX NH4
IMMOBIL. FLUX NO3
IMMOBIL. FLUX TOTN
VOLATIL. FLUX
DENIT. FLUX
NITRIFICATION FLUX
AMMONIFIC. FLUX
VAD ADVECT NH3
VAD ADVECT NO3
VAD ADVECT ORGN
VAD DISPERSION NH3
VAD DISPERSION NO3
VAD DISPERSION ORGN
VAD DECAY NH3
VAD DECAY NO3
VAD DECAY ORGN
VAD CONC NH3
VAD CONC NO3
VAD CONC ORGN
32
-------
SECTIONS
MODEL TESTING
The following section describes the model testing that was performed to evaluate the
new OSWDS and PRZM-3 codes. In Section 5.1 the study area to which the new code will
first be applied is described; to expedite this first application, study area data were used for
the code testing effort. In Section 5.2 the testing strategy is presented for modeling
components. This section focuses on specifying representative scenarios for: (1) a typical
OSWDS, (2) typical unsaturated zone flow conditions (PRZM & VADOFT zones), (3) typical
unsaturated zone nitrogen conditions (PRZM & VADOFT zones) and finally, (4) a scenario
to test uncertainty analysis related to nitrogen fate and transport beneath a septic system.
Section 5.3 presents testing procedures and results for the scenarios that were developed.
Finally, project conclusions and recommendations are provided in Section 5.4, including an
assessment of the strengths and limitations of the model enhancements.
5.1 STUDY AREA
The purpose of this section is to provide sufficient information regarding the setting
to which the the OSWDS and PRZM-3 models will be applied to enable the development of
test input sequences for the models that can be considered "typical" scenarios. To that
purpose, pertinent information has been extracted from the recent communication entitled
"Groundwater/Septic Study: North Animas Valley, La Plata, County, Colorado" (personal
communication Wright, 1995). This information includes a general description of the study
area(s), and a synopsis of the data that have been collected to support modeling efforts.
5.1.1 General Description
The following description is extracted and condensed from the above-referenced
report. Since PRZM-3 is a vadose zone model, our focus will be on defining the setting above
the water table.
The study area is directly north of Durango, Colorado, in the Upper Animas River
Valley, and is wholly contained within a glacial trough with several hundred feet of glacial,
lacustrine and alluvial deposits overlying fairly impermeable lithologies, forming a natural,
semi-confined aquifer system. In a 12-mile stretch of valley about 1.5 miles wide, three
smaller representative study areas have been defined (by EPA Region 8); in total, the study
areas encompass approximately 7.5 square miles. Within the study area(s) the Animas
River presently braids and meanders through hayfields, farms and orchards. However, this
stretch of the river is in high demand for residential development, with current population
growth rates for the county ranging from 3 to 7 percent. The greatest demand for
development occurrs in settings such as the north valley.
The groundwater levels in the study area fluctuate seasonally with variations of up
to 30 feet noted between water well levels in winter and spring. Groundwater depth tends
to increase in a southerly direction; recorded depths (Nov 1994) range from 4 feet below
ground level in the north to 110 feet below ground level in the south. There are numerous
engineered septic systems to avoid failure associated with high water and run-off. Below
the water table, the aquifer is thought to resemble a vast underground lake. It has been
33
-------
theorized that groundwater movement is limited, and consequently, that the area could be
extremely sensitive to overuse and pollution.
5.1.2 Available Data
The bulk of the data available to support this study has been compiled in a
geographic information system (GIS), that was developed at Fort Lewis College's Center for
Southwest Studies in Durango, Colorado. The coverages were created using DOS-based
pcARCINFO and UNIX-based ARCINFO. Coverages provided to the EPA are in
pcARCINFO format and pcARCINFO export format (version 3.4D) developed by Fort Lewis
College (personal communication Wright, 1995).
In addition to the coverages and attribute data in the GIS, EPA has also been
provided with long-term daily precipitation records at two nearby SCS stations, one at Fort
Lewis College and the other at Durango airport. The Fort Lewis College precipitation
record is probably more representative of the study area.
5.2 TESTING STRATEGY
Testing of both standard leaching analysis and Monte Carlo uncertainty analysis has
been performed. This section explains the rationale and selection of parameter values for
the test runs. The testing documentation in this section coincides with the input sequences
provided in attached Appendices. Modeling strategy for the OSWDS and PRZM-3 models is
divided into four components the septic system, the unsaturated zone flow, the
unsaturated zone nitrogen processes, and uncertainty analysis.
5.2.1 Typical OSWDS Scenario
The typical OSWDS scenario that we have developed assumes a three bedroom home
in which a family of four persons lives year round. The characteristics of a typical home
septic system were established using data gathered during the preparation of the
Preliminary Design Document for Phase I of this project, supported by the La Plata County
sewage disposal system regulations. The county regulations were the basis for evaluating
the depth below grade of an average septic tank outlet (and hence the depth of introduction
into the leachfield), the nature of the engineered soil column and summary of Pertinent
Data for Upper Animas Valley Study Area at the bottom of the leachfield, and a typical
leachfield size: ,
Depth below grade of sewage influx. The County regulations require
that a minimum of 12 inches of gravel enclose the sewage distribution
pipes within the leachfield trenches. The distribution pipes must be
covered by at least 2 inches of gravel, and have at least 6 inches of
gravel below them. The top, of the gravel must be covered with a layer
of hay or straw (no thickness specified). Distribution pipes are to be no
more than 36 inches below the finish grade. From these
specifications, we estimated a influx depth of 30 inches (76 cm).
Soil column above the bottom of leachfield. The regulations require
that a final cover of soil suitable for vegetation at least 10 inches deep
be placed from the top of the hay, described above, to the finished
34
-------
surface grade above the leachfield. For a system with sewage influx at
30 inches depth, a soil column (from surface downward) comprised of
approximately 20 inches (50 cm) of soil, 4 inches (10 cm) of hay, and 12
inches (30 cm) of gravel would adequately represent the engineered
soil column from the ground surface to the bottom of the leachfield
trench. Since loams and sandy loams tend to predominate in the study
area, the original soil could probably be used for fill, and hence the
model could use the characteristics of pre-installation soil for the fill
soil layer. Appropriate characterization of the hay and gravel layers
was estimated via standard references.
Typical leachfield size. The county septic systems regulations
document provides a table of minimum leachfield areas as a function
of number of bedrooms and soil percolation rate. For a three bedroom
house, the values for required leachfield area within the allowable
percolation rates range from 375 square feet to 1050 square feet. It is
likely that the loams and sandy loams that predominate the study
area would exhibit percolation rates somewhere in the mid range of
the values given. Consequently, a reasonable value for a typical
leachfield size would be approximately 650 square feet.
The following characteristics for a typical OSWDS were derived from information
contained in the Preliminary Design Document for this project:
Wastewater inflow. An average value for residential wastewater
generation is 44 gallons/person/day. For a four person household this
corresponds to 176 gallons per day. (Note: Since there are no gains or
losses to wastewater volume allowed in the OSWDS module, this value
directly determines the influx from the OSWDS module to the
appropriate layer of the PRZM-3 model. Distributed uniformly over a
650 square foot leachfield, the corresponding unit area wastewater
influx is 0.27 gallons/square foot/day, or 1.1 cm/day.)
Nitrogen species loadings in wastewater inflow. Wastewater was
characterized for the OSWDS test scenario as containing 52 mg/1 total
organic N, 11 mg/1 ammonium, and 0 mg/1 nitrate/nitrite.
r
Nitrogen species transformations in septic tank (first treatment unit).
Summaries of nitrogen species concentrations in septic tank effluent
suggest that septic tank processes result in minimal loss of nitrogen
(approximately 4% of total N is lost via septic tank pumping), with
biodegradation of organic N resulting in an approximate 3:1 ratio
between ammonium and organic N. Typically, there is negligible
nitrate-nitrite in septic tank effluent. To approximate the typical
septic tank effluent, we have assumed that the only
transformations/losses within the septic tank are from the total
organic N pool. For total organic N, the transformation factors needed
for the OSWDS module are given the following values:
transformation of total organic N to ammonium = 0.654
35
-------
loss of total organic N = 0.058
Using the above values for characterizing septic tank processes results in a
5% loss of total N and a 3:1 ratio of ammonium to organic N (45 mg/1
ammonium, 15 mg/1 organic N) in the septic tank effluent.
Nitrogen species transformations in leachfield (second treatment unit).
The second treatment unit of the OSWDS module is mainly intended
for use to estimate leachfield transformations and losses for nitrogen
species when the OSWDS module is run in a stand-alone context. In
developing the typical scenario for the linked OSWDS/PRZM-3 test
runs, soil layers that represent the leachfield will be represented in the
PRZM model; consequently, N species transformations and losses due
to denitrification and volatilization will be considered via the PRZM
model algorithms. In order to avoid "double-counting," these
transformations and losses will not be considered using the second
treatment module of the OSWDS software.
5.2.2 Typical Unsaturated Zone Flow Characteristics
Characterizing unsaturated zone flow conditions for the typical testing scenario
requires selection of a typical soil. At least three different sources of soils information are
available that could be used to provide parameter values in the PRZM-3 input sequence that
characterize flow: the EPA PATRIOT (Imhoff et al., 1994) and DBAPE (Imhoff et al., 1991)
databases, the GIS well drillers' log database, or the GIS SCS soils database. Due to our
familiarity withtheir use, the PATRIOT and DBAPE databases were combined with the GIS
for the soils information and OSWDS Nitrogen Transformation Factors for "Typical
Scenario" Test Runs. PATRIOT was also used for estimating values of hydraulic parameters
(residual water content, saturated hydraulic conductivity, van Genuchten parameters)
needed for the VADOFT simulation.
The process used to select representative soil parameter values was as follows:
(1) A search was performed on the EPA PATRIOT database to identify soils in La
Plata County; 35 soils were identified.
9
(2) A printout of the soils and areal extent of each soil was obtained from the GIS
for the study areas; 15 soils and rock outcrop were identified.
(3) Cross-checking of (1) and (2) identified eight common soils. Of these, the
PATRIOT database indicated that three had shallow layers of stratified rock
that would not enable septic system placement. The GIS database indicated
that a fourth soil constituted less than 1 percent of the study area coverage.
With the elimination of these four soils from consideration as a typical soil,
only four soils remained:
Haploborolls
Ustic Torriorthents
Shawa Variant
36
-------
Fortwingate
Preliminary examination of the characteristics from these soils suggests that any one
of them would serve adequately as the basis for evaluating the PRZM-3 model parameters
that determine flow for the preliminary model testing we performed for this project. If a
more exact selection of a typical soil is desired for subsequent model applications in the
study area, we recommend overlaying the SOILS and LEACHFLD coverages of the GIS, and
selecting the soil based on the relative location of existing leachfields.
For our purposes, the soil column of the typical case scenario was defined, starting at
the ground surface, as follows:
PRZM Model Column:
First 50 cm: soil from surface and subsurface layers, with characteristics
defined via PATRIOT soils database for Haploborolls; divided into 5 cm and
45 cm soil horizons, with only the top horizon subject to litter processes and
atmospheric deposition
Next 10 cm: layer of hay with appropriate characterization
Next 30 cm: layer of gravel with appropriate characterization
Next = 160 cm: soil from stratum (or stratum & substratum) layer(s), with
characteristics defined via PATRIOT soils database
VADOFT Model Column:
Remainder of soil column above selected "typical" water table depth of
approximately 30 feet: soil from substratum layer, with characteristics
defined via DBAPE soils database and parameter estimation capabilities
(Hayness soil)
5.2.3. Unsaturated Zone Nitrogen Characteristics
The testing strategy for soil nitrogen modeling capabilities was to develop a set of
target values for important storages and fluxes. The ability of the PRZM-3 model to produce
model results comparable to target values constitutes a successful preliminary test. The
target values were compiled from literature values and experience gained from other soil
nitrogen modeling studies. For a modeling effort that represents the influx into the soil of
nitrogen species from septic system effluent, needed information includes target values for:
Storages
Storages for N species in layers above the leachfield, below the leachfield and
within the leachfield
Belowground and aboveground plant N storages (most likely for grass)
Fluxes
N species leaching to groundwater
Plant uptake and return (grass)
Nitrification
Denitrification
Ammonia volatilization
37
-------
Mineralization
Runoff and Erosion
As noted previously, the soil N process algorithms that have been integrated into
PRZM-3 are the same as those recently integrated into the HSPF model. These algorithms
have been tested under a separate contract for a forest land modeling scenario. Since the
HSPF User's Control Input that was developed for the forest test application is the only
complete dataset (i.e., set of parameter values) currently available to execute the new
algorithms, this dataset was used as the starting point for developing the unsaturated zone
nitrogen parameter values for the PRZM-3 septic system test scenario. The surface/soil
characteristics were modified where clear differences exist between forest conditions and
the grass/subsurface septic system setting. The testing of the "typical" case septic system
scenario was refined until the code executed successfully, and the model provided
understandable and plausible results.
5.2.4 Uncertainty Analysis
The testing of the uncertainty analysis capabilities was accomplished by using the
typical case leaching analysis in a "Monte Carlo" mode, with specification of a distribution of
values for a selected PRZM-3 model parameter. Candidate parameters for uncertainty
analysis include any of the new PRZM-3 parameters, plus existing flow- and soils-related
Monte Carlo parameters (e.g., bulk density, field capacity, wilting point) for both the PRZM
and VADOFT components, as listed in the PRZM-2 manual (refer to Section 4 for further
details). Uncertainty analysis was performed on only one parameter to provide a test case
that could be more clearly interpreted. The input sequence for the Monte Carlo run is
provided in attached Appendices.
5.3 TESTING PROCEDURES AND RESULTS
The attached appendices provide the input sequence used for testing the OSWDS
module and providing 'typical case' nitrogen species fluxes from septic system effluent.
Testing of the OSWDS module was straightforward; the code successfully generated
nitrogen species fluxes for introduction into the PRZM-3 model.
As explained in Section 5.2.3, testing of the new PRZM-3 code centered on evaluating
the model's ability to reproduce target levels of nitrogen storages and fluxes; this approach
necessitates model calibration. Calibration of soil nitrogen models involves defining model
inputs, estimating the nitrogen balance expected for the soil/plant system being modeled,
and adjusting model parameters to mimic the expected or observed nitrogen balance. Most
of the soil nitrogen modeling work to date, and the majority of the currently available
literature on nitrogen balances, is based on studies of agricultural systems, with a
significantly smaller portion directed to forested systems. Frissel (1978), as reported by
Miller and Donohue (1990), shows examples of nitrogen balances developed from selected
field studies for cultivated crops, grasslands, and a few forested ecosystems. This
presentation of a nitrogen balance shows the N inputs or additions, such as
fertilizer/manure applications, N fixation, irrigation, and atmospheric deposition; the N
removals, including crop harvest, denitrification, volatilization, leaching and erosion/runoff;
and recycling processes within the soil, such as mineralization, plant uptake, and residue
return. It is important to note that the largest components of most agricultural systems are
the N additions (e.g., N applications and fixation) and resulting plant uptake and removal.
38
-------
Thus, accurately defining these two components is key to modeling soil nitrogen processes
for these systems.
Unfortunately, analogous N balance information for septic systems, like for
agricultural lands, is generally scarce. The most relevant information we have found is that
of Wall (1991) who reported on nitrogen in septic systems in Minnesota, and Brown et al.
(1984) who described a field study of nitrogen transport from three soil types under septic
fields. Wall (1991) describes (semi-quantitatively) the generation of N from a residence,
movement and transformation within the septic tank, distribution through the drainfield,
mixing with soil and other sources of N, and subsequent entry and mixing in groundwater.
Since there is little or no loss (or change) of the quantity of the N fluxes from the drainfield
to ground water, changes were made to Frissel (1978) to accommodate the expected N
balance for the soil beneath the drainfield. Thus, according to Wall (1991), the 50 ppm of
Total N leaving the drainfield is rapidly converted to mostly nitrate (49 ppm) and a small
amount remains as ammonium (1 ppm); it mixes with background nitrate mostly from
atmospheric deposition, undergoes some loss through denitrification, and then enters
groundwater in the range of 30 to 50 ppm, mostly nitrate.
i
Brown et al. (1984) performed field studies on the movement of various N species
through three different soil types under septic leach fields. They developed an approximate
N balance (Table 5.1), indicating that 2.2% of the applied N amount was lost to groundwater
from Lakeland sandy loam, 0.8% from Norwood sandy clay, and 0.4% from Miller clay; the
groundwater samples were collected 120 cm below the septic line. Although these
percentages are low, note that the application amounts for these septic systems are more
than an order of magnitude greater than the amounts added in agricultural and forested
systems. Also, the losses for both the Norwood and Miller soils were reduced due to the
slower movement of water, the greater retention of N, and the greater chance of uptake by
grass; Brown et a.l (1984) also note that the low values for the Miller soil may still be
approximating background levels and that 'breakthrough' of the applied N had not yet
occurred. The authors further mention that N was accumulating in the soil through
filtering of organics and NH4 retention, and that losses to the atmosphere (most likely
through denitrification and volatilization) could be contributing to the reduction of the N
loading from the septic lines to groundwater. For these studies the septic effluent Total N
concentration was 31.8 mg/1, about half of the value of 60 mg/1 assumed in our OSWDS
simulations. Thus, the total N reaching groundwater could be twice the amounts shown in
Table 4.0 if the N loadings are doubled.
39
-------
Table 5.1. Losses of Nitrogen (kg/ha/yr) to Groundwater from Three Soils
Beneath
Conventional Septic Systems (Brown et al., 1984)
N
Soil
Lakeland
Norwood
Miller
N03
Applied
2170
870
428
-N
NO2'-N
27.40
2.26
0.80
Organic
NH/-N
1.98
2.04
0.47
N
18.44
2.27
0.44
Total N lost to
Groundwater
47.5
6.6
1.7
Brown et al. (1984) also cited literature sources indicating that NO3-N concentrations
in the range of 20 to 30 mg/1 have been observed in groundwater below septic systems where
denitrification was limited (Stewart et al., 1979), and that 33 kg/yr (5500 kg/ha/yr) of NO3-N
can leach into groundwater below a sandy soil from a septic system serving a family of four
(Walker et al, 1973b). The range of concentrations cited is consistent with the Minnesota
study noted previously; however, the effluent flux of 5500 kg/ha/yr NO3-N is significantly
greater than other estimates we encountered in our literature search.
Based on the above literature, and past experience with soil N modeling, we
developed the approximate N balance for modeling soil systems with septic effluent
discharge presented in
Table 5.2.
40
-------
Table 5.2. Approximate N Balance for Modeling Soil Systems with Septic Effluent
Discharge
N N, kg/ha/yr
N Inputs: Septic Effluent 2400.
Atmospheric Deposition 5.
Mineralization 100.
Total N Inputs 2505.
N Outputs: Net Plant (grass) uptake 30.
Erosion and Runoff 5.
NH3 Volatilization 120.
Denitrification 800.
Leaching to Groundwater 60.
Total N Outputs 1015.
N Inputs - N Outputs: 1490.
Soil retention 1460.
Plant residue return 30.
This balance is based on six assumptions and calculations:
The septic discharge is based on 60 mg N/l for the septic effluent from a
family of four with a total discharge of 1.1 cm/day, or 0.27 gal/ft2/day for the
60.4 m2(650 ft2) leach field. (Note: A useful conversion factor is 0.1 kg/ha/day
for each 1 cm/day discharge at 1 mg/1 concentration).
Mineralization is assumed to be higher than agricultural conditions due to
the increased organics in the septic effluent.
NH3 volatilization is higher than agricultural conditions due to the high levels
of NH3 in the septic effluent. In fact, volatilization could be significantly
higher than the 120 kg/ha/day value included above.
Soil retention and denitrification are clearly the major components of the
reduction in N loadings from the septic tank discharge to the levels that reach
groundwater. Soil retention includes NH3 and organic N sorption, and
immobilization of both NH3 and NO3 (i.e. conversion to organic N).
The amount leaching to groundwater is consistent with the literature sources
reviewed, but the range is large- from < 10 to > 300 kg/ha/day, depending on
a variety of local conditions.
41
-------
The remaining components are relatively small, and the estimates are
reasonable based on past experience.
Using the nitrogen balance presented in Table 5.2 as a target, we performed a series
of eight runs of the PRZM component of PRZM-3, after which we considered the model to be
satisfactorily tested and calibrated. The input sequence for these runs is provided in the
attached Appendices, and the corresponding results of the final PRZM simulation are
presented in Table 5.2. This run represents the initial screening results for the
Haploborolls soils in the North Animas Valley.
At this point, the sample VADOFT input sequence from the current PRZM-2
distribution disk was modified to reflect the scenario described in Section 5.2 above.
Because no new code was added to VADOFT, the testing performed for the model was
limited to verifying that the model would accept the groundwater flow and nitrogen fluxes
from the PRZM component, could run successfully, and generate nitrogen fluxes at the
bottom of the vadose zone. No attempt was made to adjust model parameters related to
processes such as decay and retardation to calibrate nitrogen behavior in the vadose zone.
The VADOFT input sequence (also shown in attached Appendices) resulted in a successful
simulation of the three nitrogen species, as reflected in the results shown in Table 5.4.
As the final step in testing the code enhancements to PRZM-3, the existing Monte
Carlo code was executed to evaluate the effect on a selected nitrogen groundwater flux
(nitrate) of assigning a distribution of values to a new Monte Carlo input variable
(denitrification rate). As was the case with the VAODFT test run, the purpose of the Monte
Carlo test run was only to test the code's ability to accept new input and output variables
related to nitrogen simulation. Since no changes were made to the Monte Carlo code itself,
the test run was limited to 10 one-year runs of the PRZM component. The test run executed
successfully. The input sequence for the Monte Carlo run (including necessary changes to
the Execution Supervisor) is provided in the attached appendices, and the results of this test
run are summarized in Table 5.5.
42
-------
Table 5.3 Final Results for PRZM Component Scenario Testing
M Input:'
^ Output:
Septic Effluent
MH3
Drg. Nitrogen
Umos. Depos.
Mineralization
Total
'LANT UPTAKE
TO3
VH.3
Total
iunoff
^JH3 Vol.
Denitrification
..BACHING
M03
*H3
Drg.N
Total
Total Output
CHANGE IN N:
Starting Soil N
Ending Soil N
difference
>THER FLUXES:
'lant Return
mmobilization
*H3
403
NH3 Nitrifica.
>abile>Refrac
\G Plant Stor.
Litter N Stor.
1957
2481
1862
619
5.5
89
2576
48
12
60
0.0
114
820
8
0.0
15
23
1016
11535
13064
1529
37
452
376
1276
53
7
3
1958
2481
1862
619
5.5
89
2576
41
10
51
0.0
119
843
20
4
15
40
1053
12064
14547
1483
40
487
394
1267
60
8
3
1959
2481
1862
619
5.5
97
2584
40
10
49
0.0
125
867
24
10
16
51
1092
13547
15992
1444
44
503
408
1322
71
8
3
1960
2488
1867
621
5.5
97
2601
34
8
43
0.0
124
856
24
13
17
54
1076
15992
17446
1455
42
511
402
1312
84
9
4
1961
2481
186
619
5.5
108
2596
30
7
37
0.0
125
861
26
11
19
55
1078
17446
18894
1447
37
504
407
1321
90
9
4
1962
2481
1862
619
5.5
109
2607
16
4
20
0.0
131
902
29
11
20
59
1112
18894
20289
1395
27
530
424
1381
105
8
4
1963
2481
1862
619
5.5
120
2632
22
5
27
0.0
123
864
21
9
21
51
1064
20289
21745
1456
28
513
401
1306
128
7
4
1964
2488
1867
624
5.5
145
2646
33
8
41
0.0
129
878
27
15
23
65
1113
21745
23165
1420
35
533
414
1357
133
8
4
1965
2481
1862
619
5.5
152
2644
34
8
42
0.0
129
889
28
13
25
66
1125
24165
24569
1404
42
526
422
1350
139
8
4
1966
2481
1862
619
5.5
158
2648
18
5
23
0.0
126
882
24
11
26
61
1092
24569
25981
1412
24
516
415
1335
151
7
4
Mean
2483
1863
620
5.5
161
2587
39
10
48
0.0
121
849
21
8
16
45
1063
14517
15989
1472
40
492
397
1300
72
8
3
* Units in kg ha"1
43
-------
Table 5.4 Final Results for VADOFT Component Scenario Testing
1957 1958 1959 1960 1961 1962 1963 1964 1965 1966
Flow
Inflow 419.9 415.9 420.2 419.7 418.0 413.5 414.0 418.7 422.0 415.5
A Storage 2.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Outflow 417.2 415.9 420.2 419.7 418.0 413.5 414.0 418.7 422.0 415.5
NH3
Inflow
A Storage
Outflow
^
NO3
Inflow 81.1 202.9 243.3 239.8 256.2 285.8 211.9 271.7 278.2 260.3
A Storage 27.3 14.0 8.5 1.0 5.3 -5.5 -4.8 14.1 3.2 -6.3
Outflow 53.9 188.9 234.8 238.8 250.9 291.3 216.6 257.5 275.1 243.5
Organic N
Inflow 145.3 149.4 161.3 173.2 185.4 197.3 210.3 228.3 248.4 260.3
A Storage 75.5 3.7 6.1 6.6 6.7 7.6 6.6 9.1 9.8 8.0
Outflow 69.8 145.7 155.2 166.6 178.7 189.7 203.8 219.2 238.6 252.3
3.7
.7
3.0
42.3
10.9
31.4
102.4 127.0 109.5 106.3 88.2 144.6 129.1 111.7
3.7 3.2 -0.7 0.6 -1.0 7.7 0.5 -2.5
98.7 123.8 109.5 105.7 89.2 136.8 128.5 114.2
44
-------
Table 5.5 Results for Monte Carlo Uncertainty Analysis Test
CUMULATIVE DISTRIBUTION FOR GW FLUX NO3[1: 1]
N = 100
MEAN = 1.12
STANDARD DEVIATION = 0.891E-01
COEFFICIENT OF VARIATION = 0.797E-01
MINIMUM VALUE = 1.02
MAXIMUM VALUE = 1.30
50th PERCENTILE = 1.13 90.0 % CONF. BOUNDS = 0.000 TO 1.30
80th PERCENTILE = 1.19 90.0 % CONF. BOUNDS = 1.03 TO 1.30
90th PERCENTILE = 1.30 90.0 % CONF. BOUNDS = 1.04 TO 1.30
95th PERCENTILE = 1.30 90.0 % CONF. BOUNDS = 1.13 TO 1.30
45
-------
5.4 CONCLUSIONS AND RECOMMENDATIONS
PRZM-3 N Modeling Conclusions
1. The results over the 10-year simulation period are reasonably consistent with
the 'expected' N balance derived from review of the literature. As with the
projected N balance, the key components include the septic effluent rate as
the dominant input, denitrification and NH3 volatilization as the primary
losses, and soil retention (through adsorption and immobilization) as the
primary mechanism restricting movement to groundwater.
2. Over the first few years, the leaching of NH3 and NO3 from the soil column
shows a 'break-through' process whereby the level of leaching is low until
much of the retention capacity has been satisfied. Subsequent levels after
year three are relatively consistent throughout the remaining period.
Organic N leaching does continue to increase due to the build-up of organics
in the soil (see # 3 below)
i
3. Most other processes have very little impact due to the overwhelming loading
rate for septic systems as compared to natural or agricultural systems. For
septic systems, plant uptake/return, runoff/erosion, atmospheric deposition,
and to a lesser extent mineralization, are almost insignificant parts of the N
balance.
4. The large soil retention component, about 1500 kg/ha/yr, leads to an increase
of 125% in total soil N over the 10-year simulation period. This build-up also
impacts leaching of organic N, mineralization, and labile to refractory
conversion. Further investigation is needed to determine how closely this
matches real world conditions.
5. The calibration effort has shown that the code modifications are behaving as
expected and that, given knowledge of the relative magnitudes of the
processes actually occurring in the real world, the PRZM-3 N model can be
calibrated to mimic the system.
6. The amount leaching to groundwater, both in our expected balance and in the
model results, are only projections derived from the literature review. Local
conditions will determine to what extent the loss and soil retention processes
actually reduce the input loadings. It may be that we are greatly
underestimating the loads to groundwater, especially if there are sandy soils,
very little retention, and well aerated conditions reducing denitrification. The
actual loads could be much greater.
7. Thus, the key parameters are the rates controlling denitrification,
immobilization, and ammonia volatilization. Given conditions and an
expected N balance for a specific site, these will be the primary parameters to
be adjusted to calibrate PRZM-3.
46
-------
VADOFT N Modeling Conclusions
Since there were no enhancements made to existing VADOFT capabilities to
simulate chemical fate and transport, testing of VADOFT for this project was
limited to verifying that the nitrogen fluxes from the PRZM component could
be used successfully as input to VADOFT and that reasonable values resulted
from simulation of movement of the nitrogen species through a typical study
area soil. This level of testing was successfully performed.
Recommendations
1. The key to applying PRZM-3 for nitrogen leaching from septic systems is to
develop a reasonable expected N balance for the site conditions to be
simulated, and the key element of this N balance is the septic effluent
discharge. All the parameters that result in the discharge level need to be
evaluated for their applicability to the local conditions being represented.
2. Model users need to become knowledgeable about N fate propesses in the soil,
especially as they are affected by the septic discharge. Articles by Brown et
al. (1984), Lund et al. (1976), Walker et al. (1973a, 1973b), and Reneau (1989)
are excellent sources of such information.
3. A number of literature sources have indicated the build-up of a scum layer
near the bottom of the leach trench and a potential reduction in infiltration
due to this layer. Simulations should be performed with the restricted
drainage option used to represent this impact, and the potential impact on N
leaching.
4. Simulations need to be performed over an extended time period to evaluate
the general stability of the model results and eliminate impacts of starting
conditions.
5. Further testing of the PRZM-3 nitrogen algorithms is needed, primarily for
agricultural systems, to evaluate the plant nitrogen uptake and runoff
components. Operational testing has been performed in this effort, but
application to a specific field site is the best way to perform thorough code and
algorithm testing.
6. Additional VADOFT runs are needed to develop appropriate values for
parameters (e.g., decay, retardation) that can be used to represent the fate
and transport of ammonia, nitrate and organic N in the vadose zone.
7. PRZM-3 provides a general tool for evaluating nitrogen processes in
unsaturated soils. Joint use of the OSWDS and PRZM-3 codes provides a
more focused tool for engineers/planners who have a need to evaluate
nitrogen loadings to groundwater from septic systems. While PRZM modeling
concepts and procedures are quite familiar to professionals involved in
evaluating agricultural chemicals, the level of detail and knowledge required
for such modeling efforts is generally greater than that expended to evaluate
septic system issues. To the extent that PRZM-3 is intended for use by
47
-------
engineers/planners evaluating septic system issues, we strongly recommend
that an applications manual be developed that will guide such users through
those modeling concepts and procedures that must be resolved, and around
those that are not pertinent to septic effluent analyses.
SECTION 6
REFERENCES
Bicknell, B.R., J.C. Imhoff, J.L. Kittle, A.S. Donigian Jr., and R.C. Johanson. 1993.
Hydrological Simulation Program FORTRAN (HSPF): User's Manual for Release 10.
EPA-600/R-93/174. U. S. Environmental Protection Agency, Athens, GA.
Bicknell, B.R., J.C. Imhoff, J.L. Kittle Jr., A.S. Donigian, Jr. and R.C. Johanson. 1995.
Hydrological Simulation Program - FORTRAN. User's Manual for Release 11.
DRAFT. U.S. EPA Environmental Research Laboratory, Athens, GA.
<
Bouma, J., W.A. Ziebel, W.G. Walther, P.G. Olcott, E. McCoy and F.D. Hole. 1972.
Soil Absorption of Septic Tank Effluent. Ext. Circ. No. 20, Univ. Of Wisconsin,
Madison, WI.
Brady, N.C. 1990. The Nature and Properties of Soils. Tenth Edition. Macmillan
Publishing Co., New York, NY. 621 p.
Brooks, J.L., C.A. Rock and R.A. Struchtemeyer. 1984. Use of Peat for On-Site
Wastewater Treatment: II. Field Studies. J. Environ. Qual. 13:524-530.
Brown, K.W., K.C. Donnelly, J.C. Thomas, and J.F. Slowey. 1984. The Movement of
Nitrogen Species Through Three Soils Below Septic Fields. J. Environ. Qual.
13(3):460-465.
DeVries, J. 1972. Soil Filtration of Wastewater Effluent and the Mechanism of Pore
Clogging. J. Water Pollution Control Fed.Vol. 44:565-573.
Donigian, A.S. Jr. and W.C. Huber. 1991. Modeling of Nonpoint Source Water Quality in
Urban and Non-urban Areas. EPA-600/3-91-039. U.S. Environmental Protection
Agency, Athens, GA,
Donigian, A.S. Jr., B. R. Bicknell, and R.V. Chinnaswamy. 1995. Refinement of a
Comprehensive Watershed Water Quality Model. Draft Final Report. Prepared for
U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg, MS.
EPA. 1980. Onsite Wastewater Treatment and Disposal Systems Design Manual. Office of
Water Program Operations, Washington, DC.
Frissel, M.J.(ed). 1978. Cycling of Mineral Nutrients in Agricultural Ecosystems. Elsevier
Publishing Co., Inc. New York, NY. pp 203-243.
Hunsaker, C.T., C.T. Garten, and P.J. Mulholland. 1994. Nitrogen Outputs from Forested
48
-------
Watersheds in the Chesapeake Bay Drainage Basin. DRAFT. BSD Publication No.
4275. Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge,
TN.
Imhoff, J.C., R.F. Carsel, J.L. Kittle and P.R. Hummel. 1991. DBAPE, A Data Base and
Model Parameter Analysis System for Agricultural Soils to Support Water Quality
Management. Water Science and Technol. 24(6): 331-337.
Imhoff, J.C., P.R. Hummel, J.L. Kittle, Jr. and R.F. Carsel. 1994. PATRIOT - A
Methodology and Decision Support System for Evaluating the Leaching Potential of
Pesticides. EPA/600/S-93/010. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Athens, GA.
Kaplan, B.O. 1991. Septic Systems Handbook (Second Edition). Lewis Publishers, Inc.,
Chelsea, MI.
Knisel, W. (ed). 1980. CREAMS: A Field-Scale Model for Chemicals, Runoff, and
Erosionfrom Agricultural Management Systems. Conservation Research Report No.
26, 640 pp. U.S. Department of Agriculture, Washington, DC.
Knisel, W.G., R.A. Leonard, and F.M. Davis. 1994. GLEAMS: Version 2.10 - Part I:
Nutrient Component Documentation. Agricultural Research Service, U.S.
Department of Agriculture. Tifton, GA.
Leonard, R.A. and W.G. Knisel. 1984. Model Selection for Nonpoint Source Pollution and
Resource Conservation. In: Proc. of the International Conference on Agriculture and
Environment 1984. Venice, Italy, pp. E1-E18.
Leonard, R.A. and V.A. Ferreira. 1985. CREAMS2 - The Nutrient and Pesticide Models.
Proc. of 1983 Natural Resources Modeling Symposium. ARS-30. Agricultural
Research Service, U.S. Department of Agriculture, Washington, D.C.
Leonard, R.A., W.G. Knisel and D.A. Still. 1987. "GLEAMS: Groundwater Loading Effects of
Agricultural Management Systems." Trans, of the ASAE, 30(5):1403-1418.
Lund, L.J., A.L. Page and C.O. Nelson. 1976. Nitrogen and Phosphorus Levels in Soils
Beneath Sewage Disposal Ponds. J. Environ. Qual. 5:26-30.
Metcalf & Eddy, Inc. 1972. Wastewater Engineering: Collection, Treatment & Disposal.
McGraw-Hill Book Company, New York. 782pp.
Miller, R.W. and R.L. Donohue. 1990. Soils: An Introduction to Soils and Plant Growth. 6th
Edition. Prentice Hall, Englewood Cliffs, NJ. p 264.
Mullins, J.A., R.F. Carsel, J.E. Scarbrough, and A.M. Ivery. 1993. PRZM-2, A Model
for Predicting Pesticide Fate in the Crop Root and Unsaturated Zones: Users Manual
for Release 2.0. EPA/600/R-93/046. U.S. Environmental Protection Agency, Athens,
GA.
NVPDC. 1990. Occoquan Watershed Septic System Assessment. Submitted to the
49
-------
Virginia State Water Control Board by Northern Virginia Planning District
Commission, Annandale, VA.
NVPDC. 1992. Septic Systems Impacts in Northern Virginia: An Assessment of Two Study
Areas. Submitted to the Virginia State Water Control Board by Northern Virginia
Planning District Commission, Annandale, VA.
Reddy, K.R., R. Khaleel, M.R. Overcash, and P.W. Westerman. 1979. A Nonpoint Source
Model for Land Areas Receving"Animal Waste: II. Ammonia Volatilization. Trans.
ASAE, 22(6): 1398-1405.
Reneau, R.B. Jr., C. Hagedorn and M.J. Degen. 1989. Fate and Transport of Biological
and Inorganic Contaminants from On-Site Disposal of Domestic Wastewater. J.
Environ. Qual. 18(2):135-144.
Robertson, W.D. and J.A. Cherry. 1995. In Situ Denitrification of Septic-System Nitrate
Using Reactive Porous Media Barriers: Field Trials. Ground Water. 33(1):254-260.
i
Shaffer, M.J., A.D. Halvorson, and F.J. Pierce. (1991). Nitrate Leaching and Economic
Analysis Package (NLEAP): Model Description and Application. In: Managing
Nitrogen for Groundwater Quality and Farm Profitability. R.F. Follett, D.R. Keeney,
and R.M. Cruse (eds.). Soil Science Society of America, Inc., Madison, WI.
Stewart, L.W., B.L. Carlilie, and O.K. Cassel. 1979. An Evaluation of Alternative Simulated
Treatments of Septic Tank Effluent. J. Evniron. Qual. 8:397-403.
Tisdale, S.L., W.L. Nelson, and J.D. Beaton. 1985. Soil Fertility and Fertilizers. 4th
Edition. Macmillan Publishing Company, New York, NY. 754 p.
Wakatsuki, T. 1993. High Performance N- & P-Removal On-site Water Treatment System
by Multi-Soil Layering. Water Science and Technol. 27(l):334-345.
Walker, W.G., J. Bourma, D.R. Keeney, and F.R. Magdoff. 1973a. Nitrogen Transformations
During Subsurface Disposal of Septic Tank Effluent in Sands: I. Soil Transformation.
J. Environ. Qual. 2:475-479.
Walker, W.G., J. Bourma, D.R. Keeney, ajid P.G. Olcott. 1973b. Nitrogen Transformations
During Subsurface Disposal of Septic Tank Effluent in Sands: II. Groundwater
Quality. J. Environ. Qual. 2:521-525.
Wall. D. 1991. Septic Systems. Section I. In: Nitrogen in Minnesota Groundwater.
Prepared for the Legislative Water Commission. Minnesota Pollution Control
Agency and Minnesota Department of Agriculture, St. Paul, MN
Appendices
50
-------
Sample PRZM-3 Input
Execution Supervisor
*** PRZM2 version 2.00 | Date: Tuesday, 13 September 1994. Time: 09:27:49.***
***
* * *
*** File PRZM2.RUN, run time supervisor file for PRZM2 model, required ***
*** for all model runs. This file, as distributed by CEAM, is configured ***
*** for the data sets PRZM3.INP and VADF3.INP. Modify this file, as ***
*** as shown below within comment lines, to execute PRZM2 model with ***
*** other test input data sets distributed with PRZM2 model system. ***
*** Lines beginning with *** (i.e., three asterisks) are comment lines. ***
^^^V+H^sp^^^slc^^^^^v^ifc^&^^^^rjev^^^^^^^^^^^^^:}:^^^:^
***
*** option records
PRZM ON
VADOFT ON
*** To execute the PRZM2 model with MONTE CARLO simulation,
*** 1) turn the MONTE CARLO option switch to ON. The MONTE
*** CARLO input file that will be read from option MCIN
*** is MC.INP.
*** 2) Set the PRZM2 INPUT file option to read the files
*** PRZM3.INP and VADF3.INP.
MONTE CARLO OFF
TRANSPORT SIMULATION ON
NITROGEN SIMULATION ON
*** zone records
PRZM ZONES 1
VADOFT ZONES 1
ENDRUN
*** met stations - all on main wdm file
***WDMFILE 1 C:\PATRIOT\BIN\PRECIP.WDM
*** input file records
PATH C:\PRZM3.0\INPUT\
MCIN MC.INP
METEOROLOGY 1 W93058.MET
SEPTIC EFFLUENT 1 SEPTIC.INP
*** Change the next two lines to reflect the file names of the PRZM2
*** and/or VADOFT input files (e.g., PRZM.INP, PRZM1.INP, PRZM2.INP,
*** PRZM3.INP or VADF.INP, VADF1.INP, VADF2.INP, VADF3.INP).
PRZM INPUT 1 TESTNIT.INP
VADOFT INPUT 1 VADFNIT.INP
*** ouptut file records
PATH C:\PRZM3.0\OUTPUT\
TIME SERIES 1 TIMES.OUT
PRZM OUTPUT 1 TESTNIT.OUT
VADOFT OUTPUT 1 VADFNIT.OUT
*** MCOUT MC.OUT
51
-------
*** MCOUT2 MC2.OUT
*** scratch file records
PRZM RESTART 1
VADOFT FLOW RST 1
VADOFT TRANS RST 1
VADOFT TAPE 10 1
ENDFILES
*** global records
START DATE
END DATE
RESTART.PRZ
VFLOW.RST
VTRANS.RST
VADF.TAP
010157
311266
*** For input files PRZM.INP and PRZM1.INP the number of chemicals
*** must be set to a value of one; for PRZM2.INP and PRZM3.INP the
*** number of chemicals must be set to a value of 3.
NUMBER OF CHEMICALS 1
*** For input files PRZM.INP and PRZM1.INP comment out the next
*** two lines; for PRZM2.INP and PRZM3.INP uncomment the next
*** two lines (i.e., PARENT OF 2... and PARENT OF 3...).
*** PARENT OF 2 1
*** PARENT OF 3 2
ENDDATA
*** display records
ECHO 4
TRACE OFF
PATH C:\PRZM3.0\INPUT\
PATH C:\PRZM3.0\OUTPUT\
PATH C:\PRZM3.0\INPUT\
PATH C:\PRZM3.0\OUTPUT\
52
-------
PRZM Model
NITROGEN SIMULATION, TEMP. CORRECTION, PRZM INPUT FOR LA PLATA, CO
SEPTIC SYSTEM, SOIL N CALIBRATION RUN #7
0.75 0.44 0 15.000 1 2
0
1
1 0.20 50.0 95.000 1 58 58 58 0.0 0.0 0.0 0.0
1
020157 010557 011157 1
PESTICIDE TRANSPORT & TRANSFORMATION & APPLICATION PARAMETERS
110
No Pesticide Simulated for this run
120582 0 2.5 1.00
1 1
SOILS PARAMETERS (HAPLOBOROLLS)
250.0 0.0 000000110
O.OEO O.OEOO O.OEOO
0.60 0.60 0.45 0.16 0.25 0.25 0.25 0.25 0.25 0.25 0.18 0.16 0.97 10.0
6.0 5.0 5.0 6.0 8.0 10.0 13.0 13.0 11.0 9.0 6.0 6.0
5
5.0
0.0
1.0
5.0
45.0
0.0
5.0
7.0
10.0
0.0
2.0
7.0
30.0
0.0
5.0
7.5
160.0
0.0
20.0
8.0
0
1.27
0.25
0.0
0.0
0.0 0.000
0.25
50.0
1.27
0.12
23.0
0.25
2.32
0.0
0.0
0.0
0.0
0.0
0.0 0.000
0.25
50.0
0.20
0.12
23.0
0.05
1.16
0.0
0.0
0.0
0.0
0.0
0.0 0.000
0.05
5.0
2.2
0.0 0
0.10
30.0
1.37
0.0 0
0.20
52.0
0.01
40.0
0.0
2.0 0.0 0.0
0.10
.000
0.02
5.0
0.20
.000
0.09
22.0
0.0
0.03
0.0
0.0
0.15
0.0
0.0
0.0
0.0
0.0
0.0
0.0
53
-------
NITROGEN PARAMETERS
*** septic effluent horizon, fraction to refractory, DSNs for effluent t-series
4 .7
***vnut fora imax nupt fixn amvo alpn vnpr
0 1 100 1 0 1 1 0
*** deposition flags (AM NO3 ORGN, 3 dry, 3 wet)
-2-20000
.01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01
.005 .005 .005 .005 .005 .005 .005 .005 .005 .005 .005 .005
*** naps frmflg
0 0
*** kpln - for nupt=0
*** 0.1 0.05
*** plant uptake target and max uptake ratio (nupt=l)
60.0 2.0
*** fractions of total uptake
.013 .03 .05 .07 .13 .19 .20 .15 .085 .05 .028 .004
*** horizon fractions of uptake
.05 .05 .05 .05 .05
.85 .85 .85 .85 .85
.05 .05 .05 .05 .05
.05 .05 .05 .05 .05
.05 .05 .05 .05 .05 .05 .05
.85 .85 .85 .85 .85 .85 .85
.05 .05 .05 .05 .05 .05 .05
.05 .05 .05 .05 .05 .05 .05
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
*** above-ground plant uptake (vnut=0)
0.45 0.45 0.45 0.45 0.45
*** general parameters
***NOfr NH3fr pint
0.8 0.2 1.07 1.05
*** 1st order rates
*** ads des NOSimm
0.0 0.0 0.0 0.0005
0.0 0.0 0.0 0.0004
0.0 0.0 0.0 0.0003
0.0 0.0 0.3 0.0002
0.0 0.0 0.5 0.0002
*** max solubility
5000.0
***
des ads NOSimm min denitr nitr NHSimm
1.05 1.07 1.07 1.07 1.05 1.07
nitri NH3imm
mm
0.0
0.0
0.0
1.8
3.5
demt i
0.0
0.0
0.0
0.2
0.2
dni-tn
2.0
1.0
1.0
1.0
0.7
r m
0.5
0.5
0.2
0.5
0.5
" xfix
3.5
3.5
3.5
5.0
5.0
kf
1.0
1.0
5.0
0.5
1.0
nl
1.2
1.2
1.2
1.0
1.1
*** ammonia volatilization parameters
***theta ref T rates for each horizon
1.05 25.0 0.3 0.03 0.1 0.05 0.05
*** organic partitioning coeffs, conversion labile> refract, temp correction
1000.0 5000. 0.0002 1.07
1000.0 5000. 0.0002 1.07
1000.0 5000. 0.0002 1.07
54
-------
1000.0 5000. 0.0002 1.07
1000.0 5000. 0.0002 1.07
*** plant N return rates
*** BG return rates per horizon, fraction to refractory
0.002 0.004 0.002 0.002 0.0 0.3
*** plant>litter rate, litter>soil return rate, fraction to refractory
0.0007 0.0003 0.3
*** initial storages
***LONP LONS
RONP
300.0 0.0 1020.0
1250.0 0.0 4600.0
190.0 0.0 760.0
34.0 0.0 137.0
650.0 0.0 2280.0
*** total storages
***2424. 8797.
***AG-N Litter N
5.0 3.0
WATR YEAR
6 YEAR
RUNF TSER
STMP
STMP
STMP
0.0
0.0
0.0 45.0
0.0 8.0
0.0 50.0
263.
RONS AMSed AMSol
10.0 0.0 3.0 2.0
150.0 0.0 15.0 16.0
0.0 5.0 1.0
0.0 1.0 1.0
0.0 7.0 0.0
NO3 BG
31. 20. /total 11535
1 NITR MNTH 1 CONG YEAR
STMP
STMP
TSER
TSER
TSER
TSER
TSER
5
14
19
25
33
55
-------
VADOFT Model
AMMONIA, NO3, ORGANIC N, HAYNESS SUBSTRATUM, VADOSE ZONE FLOW
SIMULATION
71 101111100
20 2 1 .01
111101210
0.0 1.0 1.0 1.0
1 0.0 1.0
1
1 70 1 700.0
O.OOEOO 0
0 1 0.0 O.OEOO 00 0 0
1.00E01 .38EOO O.OEOO O.OEOO
0.150EOO -l.OEOO 0.026EOO 1.29EOO 0.223EOO
5 10
YEAR
AMMONIA, NO3, ORGANIC N, HAYNESS SUBSTRATUM, VADOSE TRANSPORT
SIMULATION
71 1 1 1 0 1
01100121
0.0 1.0 1.0 1.0
1 0.0 1.0
1
1 70 1 700.0
O.OEOO 1 O.OEOO 1 O.OEOO 1
00 0.0 0.0 0 0 0 0
0.12E02 .38EOO
l.OOOEOO l.OOOEOO l.OOOEOO O.OEOO O.OEOO O.OEOO
1 1.0
1 1.0
1 1.0
1 0.0 1.0 O.OEOO
1 O.OOOEOO O.OEOO
1 1
5 10
YEAR
56
-------
Sample Monte Carlo Input
Execution Supervisor
*** PRZM2 version 2.00 | Date: Tuesday, 13 September 1994. Time: 09:27:49.***
*** File PRZM2.RUN, run time supervisor file for PRZM2 model, required ***
*** for all model runs. This file, as distributed by CEAM, is configured ***
*** for the data sets PRZM3.INP and VADF3.INP. Modify this file, as ***
*** as shown below within comment lines, to execute PRZM2 model with ***
*** other test input data sets distributed with PRZM2 model system. ***
*###****************************************#*************
***
*** option records
PRZM ON
VADOFT OFF
*** To execute the PRZM2 model with MONTE CARLO simulation,
*** 1) turn the MONTE CARLO option switch to ON. The MONTE
*** CARLO input file that will be read from option MCIN
*** is MC.INP.
*** 2) Set the PRZM2 INPUT file option to read the files
*** PRZM3.INP and VADF3.INP.
MONTE CARLO ON
TRANSPORT SIMULATION OFF
NITROGEN SIMULATION ON
*** zone records
PRZM ZONES 1
***VADOFT ZONES 1
ENDRUN
*** met stations - all on main wdm file
***WDMFILE 1 C:\PATRIOT\BIN\PRECIP.WDM
*** input file records
PATH C:\PRZM3.0\INPUT\
MCIN MCNIT.INP
METEOROLOGY 1 W93058.MET
SEPTIC EFFLUENT 1 SEPTIC.INP
*** Change the next two lines to reflect the file names of the PRZM2
*** and/or VADOFT input files (e.g., PRZM.INP, PRZM1.INP, PRZM2.INP,
*** PRZM3.INP or VADF.INP, VADF1.INP, VADF2.INP, VADF3.INP).
PRZM INPUT 1 TESTNIT.INP
*** VADOFT INPUT 1 VADFNIT.INP
*** ouptut file records
PATH C:\PRZM3.0\OUTPUT\
TIME SERIES 1 TIMES.OUT
PRZM OUTPUT 1 TESTNIT.OUT
*** VADOFT OUTPUT 1 VADFNIT.OUT
MCOUT MC.OUT
MCOUT2 MC2.OUT
*** scratch file records
PRZM RESTART 1 RESTART.PRZ
57
-------
** VADOFT FLOW RST 1 VFLOW.RST
*** VADOFT TRANS RST 1 VTRANS.RST
*** VADOFT TAPE 10 1 VADF.TAP
ENDFILES
*** global records
START DATE 010157
END DATE 311257
*** For input files PRZM.INP and PRZM1.INP the number of chemicals
'*** must be set to a value of one; for PRZM2.INP and PRZM3.INP the
*** number of chemicals must be set to a value of 3.
NUMBER OF CHEMICALS 1
*** For input files PRZM.INP and PRZM1.INP comment out the next
*** two lines; for PRZM2.INP and PRZM3.INP uncomment the next
.*** two lines (i.e., PARENT OF 2... and PARENT OF 3...).
*** PARENT OF 2 1
*** PARENT OF 3 2
ENDDATA
*** display records
ECHO 4
TRACE OFF
PATH C:\PRZM3.0\INPUT\
PATH C:\PRZM3.0\OUTPUT\
PATH C:\PRZM3.0\INPUT\
PATH C:\PRZM3.0\OUTPUT\
Monte Carlo Model
MONTE CARLO INPUT FILE FOR PATRIOT RUN
10 90.0
DENITRIFICATION 4 1 0.2 0.1 0.01 0.4 1.
DENITRIFICATION 5 1 0.2 0.1 0.01 0.4 1.
END
GWFLUXN03 1 1 CDF WRITE 1
END
END
58
------- |