Modeling phosphorus in the environment - Chapter 9

Chia sẻ: bengoan741

Ô nhiễm nông nghiệp theo năm Nonpoint Nguồn (AnnAGNPS) là một mô hình công cụ đánh giá đầu nguồn công nghệ tiên tiến đã được phát triển thông qua một quan hệ đối tác giữa hai Bộ Nông nghiệp (USDA) cơ quan Hoa Kỳ - Dịch vụ Nghiên cứu Nông nghiệp (ARS) và Tài nguyên Dịch vụ Bảo tồn thiên nhiên (NRCS) - để hỗ trợ trong việc đánh giá phản ứng đầu nguồn để quản lý nông nghiệp thực hành (Bingner và Theurer 2001). AnnAGNPS là một mô phỏng liên tục, hàng ngày bước thời gian, tải gây ô nhiễm mô hình được thiết kế...

Bạn đang xem 10 trang mẫu tài liệu này, vui lòng download file gốc để xem toàn bộ.

Nội dung Text: Modeling phosphorus in the environment - Chapter 9

9 Phosphorus Modeling
in the Annualized
Agricultural Nonpoint
Source Pollution
(AnnAGNPS) Model
Yongping Yuan
U.S. Department of Agriculture-Agricultural
Research Service, Oxford, MS

Ronald L. Bingner
U.S. Department of Agriculture-Agricultural
Research Service, Oxford, MS

Indrajeet Chaubey
University of Arkansas, Fayetteville, AR


CONTENTS

9.1 Model Introduction ......................................................................................216
9.2 Watershed Processes Considered in AnnAGNPS........................................216
9.3 Model Inputs and Outputs ...........................................................................217
9.4 AnnAGNPS Model of Phosphorus Processes .............................................219
9.4.1 Soil Initial Phosphorus Content.......................................................220
9.4.2 Organic P Simulation Processes ......................................................221
9.4.3 Inorganic P Simulation Processes....................................................222
9.4.3.1 Calculation of Inorganic P Additions to a Cell ...............222
9.4.3.2 Calculation of Intermediate Inorganic
P Mass Balance..................................................................223
9.4.3.3 Calculation of Inorganic P Losses
from the Soil Profile .........................................................224
9.4.4 Total Runoff Losses .........................................................................226
9.5 Model Application........................................................................................226
9.5.1 Study Watershed and Monitoring Information................................226
9.5.2 Input Data Preparation .....................................................................227

215


© 2007 by Taylor & Francis Group, LLC
216 Modeling Phosphorus in the Environment


9.5.3 Sensitivity Analysis..........................................................................229
9.5.4 Model Calibration and Validation....................................................232
9.6 Model Limitations........................................................................................238
9.7 Conclusions ..................................................................................................238
References..............................................................................................................238


9.1 MODEL INTRODUCTION
The Annualized Agricultural Nonpoint Source Pollution (AnnAGNPS) model is an
advanced technological watershed evaluation tool that has been developed through a
partnership between two U.S. Department of Agriculture (USDA) agencies — the
Agriculture Research Service (ARS) and the Natural Resources Conservation Service
(NRCS) — to aid in the evaluation of watershed responses to agricultural management
practices (Bingner and Theurer 2001). AnnAGNPS is a continuous-simulation, daily
time-step, pollutant loading model designed to simulate long-term chemical and sedi-
ment movement from agricultural watersheds (Bingner et al. 2003). The spatial vari-
ability of soils, land use, and topography within a watershed is accounted for by dividing
the watershed into many user-specified, homogeneous, drainage-area-determined cells.
For individual cells, runoff, sediment, and pollutant loadings can be predicted from
precipitation events that include rainfall, snowmelt, and irrigation.
Each day, AnnAGNPS simulates runoff, sediment, nutrients, and pesticides
leaving the land surface and being transported through the watershed channel system
to the watershed outlet before the next day is considered. The model routes the
physical and chemical constituents from each cell into the stream network and finally
to the watershed outlet and has the capability to identify the sources of pollutants
at their origin and to track them as they move through the watershed system.
The AnnAGNPS model has evolved from the original single-event Agricultural
Nonpoint Source (AGNPS) model developed in the early 1980s by the USDA-ARS
(Young et al. 1989, 1995). The AGNPS model was developed to simulate runoff and
water-quality response of agricultural watersheds ranging from a few hectares to
20,000 hectares from a single rainfall event. The AGNPS model has been applied
throughout the world to investigate various water quality problems. The AnnAGNPS
model includes significantly more advanced features but retains many of the impor-
tant features of AGNPS. (The complete suite of AnnAGNPS model, composed of
programs, pre- and post-processors, technical documentations, and user’s manuals,
is currently available at http://www.ars.usda.gov/Research/docs.htm?docid=5199.)

9.2 WATERSHED PROCESSES CONSIDERED
IN AnnAGNPS
The hydrology components considered within AnnAGNPS are rainfall, interception,
runoff, evapotranspiration (ET), infiltration/percolation, subsurface lateral flow, and sub-
surface drainage. The runoff from each cell is calculated using the Soil Conservation
Service (SCS) curve number (CN) method (Soil Conservation Service 1985). The mod-
ified Penman equation (Jenson et al. 1990; Penman 1948) is used to calculate the potential
ET, and the actual ET is represented as a fraction of potential ET. The fraction is a linear


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 217


function of soil moisture between wilting point and field capacity. For percolation, only
the downward drainage of soil water by gravity is calculated (Bingner et al. 2003). Lateral
flow is calculated using Darcy’s equation, and subsurface drainage is calculated using
Hooghoudt’s equation (Freeze and Cherry 1979; Smedema and Rycroft 1983).
Amount of sheet and rill soil erosion loss — not field deposition — for each
runoff event is calculated using the Revised Universal Soil Loss Equation (RUSLE)
model (Renard et al. 1997). A delivery ratio, which quantifies the amount of sediment
deposited in the field and the amount of sediment delivered to the stream, is calcu-
lated using the Hydrogeomorphic Universal Soil Loss Equation (HUSLE) model
(Theurer and Clarke 1991). Ephemeral gully erosion is based on the Ephemeral
Gully Erosion model (Merkel et al. 1988). The model uses the Bagnold equation
(Bagnold 1966) to determine the sediment transport capacity of the stream and a
modified Einstein equation to determine the sediment transport in the stream system
(Bingner et al. 2003). Sediment is partitioned into five classes: clay, silt, sand, small
aggregates, and large aggregates. The model estimates particle-size distribution of
deposited sediment by taking into account the density and fall velocity of each class.
The AnnAGNPS model calculates a daily mass balance within each cell for soil
moisture, nitrogen (N), phosphorus (P), organic carbon (OC), and pesticides. Plant
uptake of nutrients, fertilization, residue decomposition, mineralization, and trans-
port are major factors considered to determine the fate of nutrients in the watershed.
Both soluble and sediment adsorbed nutrients are considered by the model.
The pesticide component is adopted from the Groundwater Loading Effects of
Agricultural Management Systems (GLEAMS) model (Leonard et al. 1987). The
AnnAGNPS model allows simulation of any number of pesticides and treats each
pesticide separately with independent equilibrium assumed for each pesticide. Both
soluble and sediment-adsorbed fractions of each pesticide are calculated on a daily
time scale. Factors affecting fate and transport of pesticides include foliage wash-
off, vertical transport in the soil profile, and degradation.


9.3 MODEL INPUTS AND OUTPUTS
A complete list of AnnAGNPS input data sections is shown in Figure 9.1. These
data can be grouped into the following categories: climate, watershed physical
information, land-management operations, chemical characteristics, and feedlot
operations. Daily precipitation, maximum and minimum temperatures, dew point
temperature, sky cover, and wind speed are climate data required by the model to
perform continuous simulation. Climate data used with AnnAGNPS can be histor-
ically measured, synthetically generated using the climate generator program
(Johnson et al. 2000), or a combination of the two.
Geographic information systems (GIS) data layers of a watershed are needed to
characterize the watershed. The GIS data layers must be in sufficient spatial detail
to permit the model to accurately reflect the real landscape it represents. Using the
GIS layers of digital elevation model (DEM), soils, and land use, a majority of the
large data input requirements can be developed using a customized ArcView GIS
interface. Those input requirements include watershed and cell delineation, cell land
slope, slope direction, cell land use and soil type, and stream reach data, can be


© 2007 by Taylor & Francis Group, LLC
218
AnnAGNPS Watershed Simulation Daily Verification Global
Identifier Data Period Climate Data Output


Field Pond
Point Feedlot
Feedlot
Gully Field Pond Management
Source Management



Cell Data Reach Data


Management
Tile Drain
Soils Field Reach Channel Reach Nutrient
Impoundment
Geometry Half Life




Modeling Phosphorus in the Environment
Management
Schedule



Runoff Curve
Fertilizer Pesticides Management Strip
Irrigation
Contours Crop
Number
Application Application Operation Crop


Fertilizer Pesticides Non-Crop
Reference Reference



Required Required if Referenced Optional


FIGURE 9.1 A complete list of AnnAGNPS input data sections.


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 219


developed by using a customized ArcView GIS interface. Additional input require-
ments, which include developing the soil layer attributes to supplement the soil
spatial layer, describing crop operations and management practices, defining channel
hydraulic characteristics, and entering many other optional data sections as needed
by the watershed (Figure 9.1), can be organized using the AnnAGNPS Input Editor.
The Input Editor is a graphical user interface developed to aid users in selecting
appropriate input parameters. Much of the information needed to characterize crop
characteristics, field operations (e.g., crop rotation, tillage, planting, harvesting),
chemical characteristics, feedlots, and soils can be obtained from databases imported
from RUSLE or from other USDA-NRCS data sources.
Feedlot information includes daily manure production rates, manure character-
istics, amount of manure removed from the field lot, and residual amount of manure
available from previous operations. The model outputs include runoff, sediment,
nutrient, and pesticide at a temporal scale ranging from daily to yearly. All model
outputs can be obtained at any desired location such as specific cells, stream reaches,
feedlots, gullies, or point sources. The model also has capabilities to provide source
accounting information in terms of the fraction of a pollutant loading passing through
any reach location that originated from a user-specified pollutant source area. Cronshey
and Theurer (1998), Geter and Theurer (1998), and Theurer and Cronshey (1998)
provide detailed information on available model outputs.



9.4 AnnAGNPS MODEL OF PHOSPHORUS PROCESSES
Simulation of P transport and transformation processes at a watershed scale is very
challenging because of the complexities and uncertainties related to the processes.
A complete understanding of the relationship of various P pools and their chemical,
physical, and biological interactions in the soil profile is essential for a full descrip-
tion of the P cycle in soils and plants (Jones et al. 1984). A model based on
mathematical descriptions of fundamental chemical, physical, and biological mech-
anisms of the soil P behavior would be ideal for P modeling.
In general, the chemical component in AnnAGNPS exists in two phases: dis-
solved (solution) in the surface runoff and attached (adsorbed) to clay-size particles
resulting from sheet and rill erosion. To simulate P loading, daily soil mass balances
of P in a cell are maintained for each computational layer. The daily mass balances
of P are adapted from the Erosion Productivity Impact Calculator (EPIC) model
(Sharpley et al. 1984; Sharpley and Williams 1990).
The P processes simulated in AnnAGNPS are shown in Figure 9.2. More
specifically, P is partitioned into inorganic P and organic P, and a separate mass
balance is maintained for each. Inorganic P is further broken down into (1) labile
P, or P readily available for plant uptake; (2) active P, or P that is more or less
reversibly adsorbed to the soil; and (3) stable P, or adsorbed P that is fixed or
relatively irreversibly chemisorbed to the soil adsorption complex or as discrete
insoluble P minerals. The model simulates the effect of P adsorption that controls P
availability to plant uptake and runoff loss, and the model also simulates P movements
between labile P and active P and between active P and stable P. Sediment-attached


© 2007 by Taylor & Francis Group, LLC
220 Modeling Phosphorus in the Environment


Inorganic Organic
Organic
Inorganic Erosion
Plant
Runoff
Erosion Plant residue
fertilizer
uptake fertilizer loss
loss
loss




Mineralization
Desorption
Active and
Active Solution
Stable
Decay
(Humic)
Adsorption


Residue mineralization Fresh
Stable




FIGURE 9.2 Phosphorus processes simulated in AnnAGNPS.



P estimated from soil erosion is assumed to be associated with the clay-size
fraction of the soil and consists of both organic and inorganic P. Major processes
considered are residue decomposition and mineralization, fertilizer application,
plant uptake, runoff, and erosion losses. Plant uptake of P is modeled through a
simple crop-growth stage index either specified by the user or by the model
(Bingner et al. 2003).
Phosphorus losses from each AnnAGNPS cell within a stream reach are added
to an AnnAGNPS reach. Phosphorus is reequilibrated between dissolved P and
sediment-attached P in the reach during transport to the watershed outlet.

9.4.1 SOIL INITIAL PHOSPHORUS CONTENT
The initial soil P content is needed to initialize AnnAGNPS simulation. Usually,
calibration is recommended to define the initial soil P content.
The input P levels in the soil profile are input as concentrations, but AnnAGNPS
performs calculations on a mass basis. To convert a concentration to a mass, AnnAGNPS
uses a conversion factor, conv (Equation 9.1). The conversion factor converts nutrient
concentration in soil to mass (in kilograms) using Equation 9.1:

conv = 10,000 ρ b D Acell (9.1)

where conv is the intensive unit-to-extensive-unit conversion factor (kg), ρb is the
bulk density of composite soil layer (g/cm3 or mg/ m3), D is thickness of soil layer
(mm), and Acell is the AnnAGNPS cell area (ha).



© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 221


9.4.2 ORGANIC P SIMULATION PROCESSES
All AnnAGNPS mass balances are based on AnnAGNPS cells and are maintained
for two composite soil layers. The first soil layer is 203 mm in depth from the
surface, typically defined as the tillage layer by RUSLE. The second soil layer is
from the bottom of the tillage layer to either an impervious layer or the user-supplied
depth of the soil profile.
The mass balance equation for organic P simulation is as follows:

(resP + orgPfer − hmnP − orgPsed ) 1,000,000
orgPt = orgPt −1 + (9.2)
conv
where orgPt is organic P concentration in the composite soil layer for the current
day (mg/kg), orgPt–1 is organic P concentration in the composite soil layer for the
previous day (mg/kg), resP is organic P addition to a cell from decomposed fresh
crop residue (kg), orgPfer is organic P addition to a cell from fertilizer application
(kg), hmnP is the mineralization from the humus active organic P pool (kg), and
orgPsed is organic P loss from a cell by attaching to sediment (kg).
Decomposition is calculated once a day. Equations for residue decomposition
were adapted from RUSLE. Only surface decomposition is calculated for crop land.
Cell organic P from fertilizer application is the product of the fertilizer applied for
the current day and the organic P fraction in the fertilizer. The organic P fraction
can be obtained from the fertilizer reference database in AnnAGNPS.
The P mineralization equation is adapted from the EPIC model (Sharpley and
Williams 1990). Temperature and aeration, represented by soil moisture, are con-
sidered for P mineralization (Sharpley and Williams 1990). AnnAGNPS assumes
that organic phosphorous is associated with the clay fraction of the soil. Sediment-
attached organic P is calculated by Equation 9.3:

orgPsed = forgP × sed clay × 1000 (9.3)

where forgP is a decimal fraction of organic P in clay in soil layer (g/g), and sedclay is
the amount of clay in the mass of sediment (mg). The decimal fraction of organic P is:

orgP
forgP = (9.4)
fclay × 1,000,000

where orgP is the organic P concentration in the composite soil layer (mg/kg), and
fclay is the fraction of clay to total composite soil, provided by the soil database.
Organic P mass balance is maintained for the second soil layer the same way
as the first layer except that fertilizer application and rainfall-induced runoff and
sediment loss are not considered. AnnAGNPS assumes that fertilizer application,
rainfall-induced runoff, and sediment loss are associated only with the top soil layer.
Equation 9.5 represents the mass balance for the second layer:

hmnP 1,000,000
orgPt = orgPt −1 − (9.5)
conv



© 2007 by Taylor & Francis Group, LLC
222 Modeling Phosphorus in the Environment


9.4.3 INORGANIC P SIMULATION PROCESSES
AnnAGNPS simulates three different pools of inorganic P in the soil. It adapts the
principles of the soil mineral P model developed by Jones et al. (1984). Mineral P
is transferred among three forms: labile P in solution (available for plant use and
runoff loss), active P, and stable P. AnnAGNPS assumes that inorganic P added from
fertilizers initially goes to the labile P pool and the active P pool, based on a value
of the P sorption coefficient. Fertilizer P that is labile at application may be quickly
transferred to the active mineral pool. Many studies have shown that after an appli-
cation of inorganic P fertilizer, solution P concentration in the soil decreases rapidly
with time due to reaction with the soil. This initial fast reaction is followed by a
much slower decrease in solution P that may continue for several years (Barrow and
Shaw 1975; Munns and Fox 1976; Rajan and Fox 1972; Sharpley 1982). Flow
between the active and stable mineral pools is governed by a P exchange rate.
Within each inorganic P pool, addition from fertilizer application is calculated
first, followed by the mineralization of organic P. Then, losses through runoff,
erosion, and plant uptake are calculated. At the end of each day, the mass balance
is updated for each P pool. The simulation is a sequence of adjusting the mass
balance of each inorganic P pool.

9.4.3.1 Calculation of Inorganic P Additions to a Cell

Fertilizer additions are simulated in one of two ways: well mixed with the top soil
layer or unincorporated on the soil surface. On a daily basis, AnnAGNPS checks if
there is a tillage operation and the percentage of soil disturbance from the tillage
operation. If the soil disturbance exceeds 50% of the top soil layer, any fertilizer
applications are considered as mixed. Otherwise, it assumes the applied fertilizer
stays on the soil surface. In addition, when the soil disturbance exceeds 50% of the
soil, it incorporates not only the applied fertilizer on the current day but also any
fertilizer left on the soil surface from previous applications. Therefore, when soil
disturbance exceeds 50% of the top soil layer,
mnaP = surf _ inorgP (9.6)
where mnaP is the mass of inorganic P added to the soil profile from the current operation
(kg) (and it is assumed to be well mixed with the first soil layer), and surf_inorgP is the
surface inorganic P in a cell, added through fertilization at the soil surface (kg).
If a fertilizer is applied in the current operation, then

mnaP = mnaP + inorgPfer (9.7)
where inorgPfer is inorganic P applied during the current operation (kg). It is calcu-
lated using the rate of fertilizer applied for the current day times the inorganic P
fraction (from the fertilizer reference database mass/mass).
When soil disturbance is less than 50% of the soil, the fertilizer on the soil
surface remains on the soil surface and nothing is incorporated into the soil profile.
If a fertilizer is applied for the current operation, then

surf _ inorgP = surf _ inorgP + inorgPfer (9.8)


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 223


Then, AnnAGNPS checks if a rainfall event occurred, and if so, soil inorganic
P is adjusted to reflect the rainfall impact. When a rainfall event occurs, it dissolves
the soluble P on the soil surface. When the rainfall generates runoff, AnnAGNPS
assumes that inorganic P on the soil surface is totally dissolved in the water and is
either carried away with runoff or is carried into the soil profile with infiltration.
The amount of inorganic P carried away with runoff or carried into the soil profile
with infiltration is determined based on the amount of runoff and infiltration from
the rainfall event.


Q
surf _ sol _ P = surf _ inorgP (9.9)
(Q + inf )

inf
inf _ sol_ P = surf _ inorgP (9.10)
(Q + inf )

where surf_sol_P is mass of inorganic P in runoff (kg), inf_sol_P is the amount of
inorganic P carried into the soil profile by infiltration (kg), Q is the amount of surface
runoff (mm), and inf is the amount of infiltration (mm). Then, the amount of inorganic
P carried into the soil profile by infiltration is added to the mnaP value to reflect
the impact of the current rainfall event.


9.4.3.2 Calculation of Intermediate Inorganic P Mass Balance
The intermediate inorganic P mass balance refers to P pools with P additions but
prior to any P losses to runoff, erosion, and plant uptake. Bottom soil-layer inorganic
P does not change with this operation.
A portion of the incorporated inorganic P is added into the labile P pool:


Psp mnaP 1,000,000
labPi = labPstart − mpr + (9.11)
conv

where labPi is the concentration of intermediate labile inorganic P in the composite
soil layer (mg/kg), labPstart is the concentration of labile inorganic P at the beginning
of a day, and it is equal to the labile P at the end of the previous day (mg/kg), mpr
is the flow rate of P between labile and active P pools on the current day (+ implies
flow from labile to active pool; – implies flow in the opposite direction) (mg/kg/d)
(Sharpley and Williams 1990), Psp is the soil type-dependent P sorption coefficient
(dimensionless) (Sharpley and Williams 1990), and mnaP is mass of inorganic P
added to a cell soil profile (kg).
The rest of the incorporated inorganic P is added into the active P pool:

(1 − Psp) mnaP 1,000,000
actPi = actPstart + mpr + − aspr (9.12)
conv


© 2007 by Taylor & Francis Group, LLC
224 Modeling Phosphorus in the Environment


where actPi is the concentration of intermediate active inorganic P in the composite
soil layer (mg/kg), actPstart is the concentration of active inorganic P at the beginning
of a day (equal to the active P at the end of the previous day) (mg/kg), and aspr is
the flow rate of P between active and stable P pools on the current day (+ implies
flow from active to stable pool; – implies flow in the opposite direction) (mg/kg/d)
(Sharpley and Williams 1990).
Stable P pool size is calculated as follows:


stbPi = stbPstart + aspr (9.13)

where stbPi is concentration of intermediate stable inorganic P in the composite soil
layer (mg/kg) and stbPstart is the concentration of stable inorganic P at the beginning
of a day (equals to the stable P at the end of the previous day) (mg/kg).
Then, the inorganic P pools are further adjusted to add the organic P from
mineralization. This mineralized P is partitioned among three inorganic P pools
based on the fraction of each inorganic P pool to total inorganic P.

hmnP flab 1,000,000
labPi+1 = labPi + (9.14)
conv

hmnP × fact 1,000,000
actPi+1 = actPi + (9.15)
conv

hmnP × fstb 1,000,000
stbPi+1 = stbPi + (9.16)
conv

where hmnP is the mineralization from the humus active organic P pool in the soil
layer on the current day (kg), flab is the fraction of labile P to total P (total P is the
sum of labile P, active P, and stable P), fact is the fraction of active P to total P, and
fstb is the fraction of stable P to total P.

9.4.3.3 Calculation of Inorganic P Losses from the Soil Profile

This calculation includes sequential adjustments to the P pool size to reflect losses
from a cell.

9.4.3.3.1 Loss through Surface Runoff
When a rainfall event occurs, runoff interacts with soil and carries soluble inorganic
P in the soil profile away from fields. AnnAGNPS assumes the effective depth of
runoff interaction with soil to be 10 mm. All soluble inorganic P in the top 10 mm
of soil is carried away by the runoff.
Soil soluble inorganic P in the top soil layer available for runoff loss is calculated as

labP
soil_ sol_ inorgP = (9.17)
(1 + Kd _ inorgP)


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 225


where soil_sol_inorgP is the concentration of soluble P available for runoff loss in
a cell soil profile on the current day (mg/kg) and Kd_inorgP is the linear partitioning
coefficient for inorganic P (the ratio of the mass of adsorbed P to the mass of P in
solution).
Soluble inorganic P removed by runoff from the top 10 mm of soil is calculated as

soil _ sol _ inorgP Conv
cell_ soil_ sol_ inorgP = edi (9.18)
Depth 1,000,000

where cell_soil_sol_inorgP is the inorganic P removed from the top soil layer
through runoff (kg), edi is the effective depth of interaction factor, AnnAGNPS uses
10 mm, and depth is the depth of the top soil layer (mm).
The labile P pool is adjusted to reflect the loss to surface runoff.

cell_ soil_ sol_ inorgP 1,000,000
labPi+ 2 = labPi+1 − (9.19)
conv

9.4.3.3.2 Loss to Soil Erosion
Soil erosion also carries inorganic P away from fields. The inorganic P loss through
erosion is calculated the same way as organic P. AnnAGNPS assumes that the
inorganic P is also associated with clay fraction. The amount of sediment-attached
inorganic P is calculated first; then it is partitioned between the active and stable P
pools based on the amount of each pool.

sed _ inorgP _ actP 1,000,000
actPi+ 2 = actPi+1 − (9.20)
conv

sed _ inorgP _ stbP 1,000,000
stbPi+ 2 = stbPi+1 − (9.21)
conv
where sed_inorg_actP is the sediment loss from active P (kg) and sed_inorg_stbP
is the sediment loss from stable P (kg).

9.4.3.3.3 Loss through Plant Uptake of Inorganic P
In AnnAGNPS, the amount of crop nutrient uptake is calculated in a crop-growth
stage subroutine that determines the crop-growth stage based on crop data specified
by a user. Four growth stages —initial, development, mature, and senescence — are
simulated by AnnAGNPS. The length of each growth stage can be specified by a
user or by the model (Bingner et al. 2003). The amount of nutrient uptake is
calculated based on the crop-growth stage and differs by growth stage. The crop
nutrient uptake is also limited by available nutrients in the composite soil layer. The
calculated crop uptake P in the crop-growth stage subroutine affects the inorganic
P mass balance. Phosphorus uptake on a given day is calculated as follows:

growth_ P_ uptake yield P _ uptake_ harvest
uptP = Acell (9.22)
stage_ length


© 2007 by Taylor & Francis Group, LLC
226 Modeling Phosphorus in the Environment


where uptP is the amount of inorganic P taken up by the plant on the current day (kg),
growth_P_uptake is the fraction of P uptake for the current growth stage, yield is
the yield at harvest (kg/ha), P_uptake_harvest = P uptake per yield unit at harvest
(mass of P/mass of harvest unit, dimensionless), and stage_length is the the number
of growing days for the current growth stage (days).
Plant P uptake is adjusted based on the availability of P in the soil. If uptP
calculated in Equation 9.22 is greater than the available labile P in the soil layer,
then a limited crop P uptake is calculated as

labPi+ 2 conv
uptPlimited = 0.99 (9.23)
1,000,000
where uptPlimited is the mass of labile P taken up by the plant on the current day (kg)
and labPi+2 is the labile P concentration in the soil (mg/kg). The mass of crop uptake
P is subtracted from the labile P pool at the end of each day.

9.4.4 TOTAL RUNOFF LOSSES
The total mass of inorganic P lost through surface runoff is composed of loss from
the soil profile (Equation 9.18) and loss from the soil surface (Equation 9.9). Due to
the low mobility of P, leaching loss of soluble P is not simulated. Phosphorus losses
from each AnnAGNPS cell within a stream reach are added to an AnnAGNPS reach.
Phosphorus is reequilibrated between dissolved P and sediment-attached P in the reach
based on the P partitioning coefficient during the process of being transported to the
watershed outlet. Detailed P transformation in the reach is not simulated.


9.5 MODEL APPLICATION
AnnAGNPS is currently utilized in many locations of the U.S. by the Environmental
Protection Agency (EPA), NRCS, and others to estimate the impact of best man-
agement practices on nonpoint pollution (Yuan et al. 2002). Several studies have
been performed to evaluate the performance of AnnAGNPS in predicting runoff,
sediment, and nitrogen losses (Baginska et al. 2003; Suttles et al. 2003; Yuan et al.
2001, 2003). Suttles et al. (2003) evaluated AnnAGNPS performance on P simulation
in a coastal plain agricultural watershed in Georgia, and Baginska et al. (2003)
performed a similar evaluation on a small experimental catchment in the Sydney
region of Australia. This section presents the AnnAGNPS application to the Deep
Hollow (DH) watershed and evaluates the performance of AnnAGNPS on P simulation
using comparisons with measurements from the DH watershed of the Mississippi Delta
Management Systems Evaluation Area project (MDMSEA).

9.5.1 STUDY WATERSHED MONITORING INFORMATION
AND

Data collected at the DH watershed by Yuan et al. (2001) were used to evaluate the
performance of the AnnAGNPS P component. The DH watershed, located in Leflore
County, Mississippi, is one of three watersheds studied in the MDMSEA, which
seeks to develop and assess alternative innovative farming systems for improved


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 227


water quality and ecology in the Mississippi Delta. The main crops grown in the
DH watershed are cotton and soybeans. The watershed contains 15 soil series varying
in texture from loamy sand to silty clay, but three series cover 80% of the total area
(Yuan et al. 2001). Detailed records of agricultural operations including tillage,
planting, harvesting, fertilization, cover crop planting, and pesticide usages have
been maintained since 1996 (Yuan et al. 2001). A rate of 72.9 kg/ha phosphate
fertilizer was applied to cotton fields on October 6, 1998, with equipment that knifes
in the material at a depth of 100 mm without further mixing with soil. No fertilizer
was applied to soybean fields or during the winter wheat cover crop-growth period.
In the period of 1995 to 1996, the U.S. Geological Survey (USGS) installed a
gauging station to monitor runoff, sediment, nutrient, and pesticide loadings at one
of the inlets to the DH Lake (Yuan et al. 2001). Data collected at this monitoring
site were used for this study. The drainage area for the monitored site was 11 ha.
Runoff was monitored using a critical flow flume. Both discrete and composite
samples were taken during rainfall events for sediment and nutrient analyses. Rainfall
was monitored at the flume using a tipping bucket rain gauge.
Total P and orthophosphate concentrations were determined for water samples.
Total P and orthophosphate mass loads were calculated by using discrete samples
when available (Rebich 2004). Loads were also calculated by using composite
samples for runoff events when discrete samples were not available (Rebich 2004).

9.5.2 INPUT DATA PREPARATION
Established input files for model runoff and sediment evaluation — watershed topo-
graphy, soil type, climate data, and actual field operations and management (Yuan
et al. 2001) — were modified for this study. Yuan et al. (2001) described the
development of input information for AnnAGNPS simulations (complete informa-
tion on input file preparation can be found at the AGNPS website at
http://www.ars.usda.gov/ Research/docs.htm?docid=5199). The subwatersheds
(AnnAGNPS cells), land use, soil information, and stream network for the monitoring
site are presented in Figure 9.3. Based on this input file, fertilizer application was
timed according to actual field records. Fertilizer application reference information
was set up based on AnnAGNPS guidelines and databases.
Detailed soil information was obtained from the Soil Survey Geographic
(SSURGO) Database (Natural Resources Conservation Service 2005). SSURGO
provides most of the soil parameters needed for AnnAGNPS simulation, such as
soil texture, erosive factor, hydraulic properties, pH value, and organic matter.
However, information on soil nutrient contents was not available from this database.
Determining initial soil nutrient values needed for the model was a very difficult
task. Soil testing is one way of gaining soil nutrient values. Location, timing, and
method of sampling impact the nutrient values obtained from soil testing (Self and
Soltanpour 2004). However, soil testing may not be a feasible way to gain soil
nutrient values at a watershed scale because of limited resources. First, a watershed
may include thousands of fields. Second, each field has different soil types and field
managements. Third, nutrient level may vary from one location to another within a
field. Consequently, obtaining representative values for the watershed is challenging.


© 2007 by Taylor & Francis Group, LLC
228 Modeling Phosphorus in the Environment


Stream location


41
42
Gauging station
43
32
52
51
33 53
62
61
22
63

23




Cell ID Area Soil Type Hydrologic Land Use
(ha.) Soil Group
22 0.62 284B Tensas silty clay loam D Soybeans
23 2.2 284B Tensas silty clay loam D Cotton
32 0.32 284B Tensas silty clay loam D Soybeans
33 0.39 284B Tensas silty clay loam D Soybeans
41 2.85 178A Dundee loam C Cotton
42 0.53 284B Tensas silty clay loam D Soybeans
43 0.94 284B Tensas silty clay loam D Cotton
51 0.53 164B Dubbs very fine sandy loam B Cotton
52 0.12 12A Alligator clay D Cotton
53 0.59 178A Dundee loam C Cotton
61 0.81 178A Dundee loam C Cotton
62 0.24 12A Alligator clay C Soybeans
63 1.15 284B Tensas silty clay loam D Cotton


FIGURE 9.3 Subwatersheds (cells), land use, soil information, and stream network for the
monitoring site. (From Y. Yuan, R.L. Bingner, and R.A. Rebich, Trans. ASAE 44(5),
1183–1190, 2001. With permission.)



Therefore, a sensitivity analysis is needed to identify how initial soil nutrient levels
impact the simulation result. Literature searches have found that total P in surface
soils ranges from 50 to 1500 mg/kg and decreases with depth (Havlin et al. 1999).
Organic P typically varies between 15 and 90% of the total P in soils. Thus, as base
values for sensitivity analysis, initial soil organic P content was set to 500 mg/kg
for the top layer and 250 mg/kg for the subsequent layers; initial soil inorganic P
content was set to 250 mg/kg for all soil layers.
Detailed crop information such as crop yield, growth period, and amount of
residue produced was imported from the RUSLE crop database. However, plant
nutrient uptake information was not available from this database. Determining plant


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 229


nutrient uptake for this study was another challenge because information on plant
nutrient uptake is usually not available at a watershed scale. AnnAGNPS requires
plant nutrient uptake, which is expressed as a ratio (weight of P/weight of dry matter
at harvest), through the crop data section. AnnAGNPS converts the plant P uptake
value into a daily value based on crop-growth stage. To gain information on plant
nutrient uptake, an intensive literature search was conducted. Research on various
cottons in Alabama and Louisiana showed that an average of 58 kg N/ha and 9.1
kg P/ha was removed when seed cotton was harvested under optimum fertilization
condition (Bassett et al. 1970; Boquct and Breitenbeck 2000; Mullins and Burmester
1990). Based on the total dry matter production of cotton at harvest (Bassett et al.
1970; Boquct and Breitenbeck 2000; Mullins and Burmester 1990), cotton N uptake
was set at 0.017, and cotton P uptake was set at 0.0023 (Boquct and Breitenbeck
2000; Mullins and Burmester 1990). Similarly, soybean N uptake was set at 0.092
and soybean P uptake at 0.0095 (Flannery 1986), whereas winter wheat N uptake
was set at 0.022 and winter wheat P uptake at 0.0025 (Baethgen and Alley 1989).
Taking these as base values, a sensitivity analysis is also needed to identify how
plant P uptake impacts the simulation result.

9.5.3 SENSITIVITY ANALYSIS
The purpose of a sensitivity analysis is to investigate input parameters, especially
those that are difficult to measure or whose expected effect on model output is
unclear (Lane and Ferreira 1980). The purpose of this sensitivity analysis was to
identify parameters with effects that were greatest on P losses so that model users
could focus their data collection on the more sensitive parameters.
In a study of the Water Erosion Prediction Project (WEPP) model sensitivity,
Nearing et al. (1990) used a single value to represent sensitivity of the output
parameter over the entire range of the input parameter tested. The index described
by Equation 9.24 (Nearing et al. 1990) was selected for sensitivity testing of the
AnnAGNPS P losses.

O2 − O1
O12
S= (9.24)
I 2 − I1
I12

where I1 and I2 are the least and greatest values of input used, respectively, I12 is the
average of I1 and I2, O1 and O2 are the output values in response to the two input
values, and O12 is the average of O1 and O2.
The parameter S represents the ratio of a relative normalized change in output
to a normalized change in input. An index of 1 indicates a one-to-one relationship
between the input and the output, such that a 1% relative change in the input leads
to a 1% relative change in the output. A negative value indicates that input and
output are inversely related. The greater the absolute value of the index, the greater
the impact that an input parameter has on a particular output. Because it is dimen-
sionless, S provides a basis for comparison among input variables.


© 2007 by Taylor & Francis Group, LLC
230 Modeling Phosphorus in the Environment



TABLE 9.1
Input Parameters Considered in the Sensitivity Analysis
Values
Input Parameters A B(Base Value) C

P Mixing Code YES NO NA
P application rate (kg/ha) NA 72.9 353.0
Initial soil P Organic P 50 500 NA
content in Inorganic P 25 250 NA
the top soil
layer (mg/kg)
Plant P uptake Cotton 0.0003 0.0023 0.0043
(ratio) Soybean 0.0075 0.0095 0.0115
Winter wheat 0.0005 0.0025 0.0045

= no value selected for sensitivity analysis for that situation.
Note: NA




As noted previously, soil nutrient content and plant nutrient uptake are difficult
parameters to measure. Because the impact of fertilizer application on nutrient losses
is a public concern, sensitivity analysis was performed for inorganic fertilizer appli-
cation. The study investigated the sensitivity of the P losses from a watershed to
changes in the following input parameters: (1) P mixing code, (2) P application rate,
(3) initial P content in the top soil layer, and (4) plant P uptake. Phosphorus mixing
code reflects how well the applied fertilizer is mixed within the depth of application.
If the P mixing code is set to “YES,” applied fertilizer is well mixed within the
depth of application. If the P mixing code is set to “NO,” the entire applied fertilizer
remains on the soil surface.
Values used for sensitivity analysis are listed in Table 9.1. For the P application
rate, the actual amount of fertilizer applied was used as the base value. Base-value
selections for initial soil P content and plant P uptake were discussed in Section 9.5.2.
Each parameter varied individually within a range as reported in the literature
(Baethgen and Alley 1989; Bassett et al. 1970; Boquct and Breitenbeck 2000;
Mullins and Burmester 1990).
The AnnAGNPS sensitivity analysis simulation was performed over a 4-year
period. Annual average P loss was used as the output parameter for the sensitivity
analysis. To evaluate the sensitivity of P mixing code (Table 9.2), four AnnAGNPS
simulations were performed: two for P application rate B and two for P application
rate C. Base values listed in Table 9.1 were used for initial soil P content and plant
P uptake for all four simulations. No sensitivity index can be calculated for this
analysis because no quantitative number is associated with P mixing code “YES”
or “NO.” A percent error, which indicates relative changes of P losses from P mixing
code “YES” to “NO,” was calculated (Table 9.2). First, the difference between P
losses from P mixing codes “YES” and “NO” was calculated, and then the percent
error was calculated as a ratio between the difference and P loss from P mixing code


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 231



TABLE 9.2
Sensitivity Analysis for the Effect of Mixing Code on P Losses
Annual Average P Losses (kg/ha)
P Application Rate B P Application Rate C
P Mixing Code Attached P Dissolved P Attached P Dissolved P

YES 1.683 15.610 1.707 18.713
NO 1.680 16.514 1.695 22.702
−0.055 −0.176
Percent error (ratio) 0.001 0.007

Notes: Application rate B = 72.9 kg/ha; application rate C = 353.0 kg/ha. Initial soil P
content and plant P uptake remain as base values listed in Table 9.1. The percent error
is calculated as the ratio of the difference between results from P mixing code “YES”
and “NO” and results from P mixing code “NO.”




TABLE 9.3
Sensitivity Indexes, S, of Selected Parameters on P Losses
Sensitivity Index (S) for Sensitivity Index (S) for
P Mixing Code “YES” P Mixing Code “NO”
Input Parameters Attached P Dissolved P Attached P Dissolved P

P application rate (kg/ha) 72.9 0.01 0.14 0.01 0.24
353.0
Initial soil Organic P 500 0.54 0.05 0.57 0.05
P content 50
in the top Inorganic P 250 0.19 0.94 0.22 0.84
soil layer 25
(mg/kg)
Plant P Cotton 0.0003 –0.003 –0.03 –0.003 –0.03
uptake 0.0043
(ratio) Soybean 0.0075 –0.014 –0.12 –0.013 –0.11
0.0115
Winter 0.0005 –0.004 –0.03 –0.004 –0.03
wheat 0.0045




“NO.” The impact of P application rate on model predictions (Table 9.3) was
performed using P application rates of 72.9 kg/ha and 353 kg/ha, and base values
listed in Table 9.1 were used for initial soil P content and plant P uptake. The
sensitivity index, S, was calculated using Equation 9.24 for sediment-attached P and
dissolved P (Table 9.3). Similarly, the impact of initial soil P content and plant P
uptake on model predictions was performed (Table 9.3).


© 2007 by Taylor & Francis Group, LLC
232 Modeling Phosphorus in the Environment


Sensitivity analysis results (Table 9.3) indicate that the most sensitive variables
of those selected for analysis of P losses were initial soil P contents. This is consistent
with previous studies (e.g., Fang et al. 2002; Pote et al. 1996, 1999; Sharpley 1995),
which demonstrated that P losses to surface runoff were significantly correlated with
soil P levels. Attached, or particulate, P loss is more sensitive to the initial soil
organic P than soil inorganic P, but is not sensitive to fertilizer application rate and
plant P uptake (Table 9.3). In contrast, dissolved P loss is very sensitive to the initial
soil inorganic P, less sensitive to P application rate, and not sensitive to plant P
uptake (Table 9.3). Because no sensitivity index can be calculated for the P mixing
code, the sensitivity of P losses to P mixing code differs from the other parameters
analyzed. It is similar, however, to P application rate in that the dissolved P is
sensitive to the P mixing code whereas the attached P is not. The sensitivity of P
losses to the P mixing code increases with the increase of P application rate as
expected (Table 9.2).

9.5.4 MODEL CALIBRATION VALIDATION
AND

Since initial soil P content had the greatest impact on P losses among parameters
tested for sensitivity (Table 9.3), the initial soil P content was adjusted to give good
correspondence with the observed P losses. Initial soil P content selection for
calibration was based on many studies of P summarized in Havlin et al. (1999). The
first simulation was performed using 100 mg/kg for initial soil organic P content
and 15 mg/kg for initial soil inorganic P content, which represents the lower level
of soil P content. The second simulation was performed using 500 mg/kg for initial
soil organic P content and 250 mg/kg for initial soil inorganic P content, which
represents the average level of soil P content. The third simulation was performed
using 1000 mg/kg for initial soil organic P content and 500 mg/kg for initial soil
inorganic P content, which represents the high level of soil P content. The selection
of initial soil P contents involved many trials and errors. The first 27 months were
used to calibrate the model, and the last 22 months were used to validate the model.
For plant P uptake, base values used for sensitivity analysis (Table 9.1) were chosen
because these values were typical values under optimum fertilization. It is assumed
that fertilizer applied in this study was the optimum value for crop uptake. Ann-
AGNPS predicts P loss in dissolved and sediment-attached phases; thus, the pre-
dicted total P loss was generated by summing dissolved and sediment-attached P
losses (Table 9.4). The predicted and observed P losses listed in Table 9.4 do not
include all P losses from the watershed. Although an attempt was made to collect
samples for every storm event, some storm events were not sampled due to unfore-
seen circumstances such as equipment malfunctions. Therefore, comparisons
between model simulations and observations were made only when monitoring data
were available. Linear regression and Nash-Sutcliffe coefficient of efficiency (Nash
and Sutcliffe 1970) were calculated to evaluate the model’s performance (Table 9.4).
The Nash-Sutcliffe coefficient of efficiency, E, ranges from −∞ to 1, with 1 indicating
the model is a perfect prediction (Nash and Sutcliffe 1970).
Calibration results showed that AnnAGNPS underpredicted total P loss and
overpredicted dissolved P loss. Calibration demonstrated that increases in either



© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling
TABLE 9.4
Monthly Observed Rainfall, Observed and Predicted Runoff, Sediment Loss, Dissolved P, and Total P Losses

Runoff (mm) Sediment Loss (mg/ha) Dissolved P Loss (g/ha) Total P Loss (g/ha)
Rainfalla
(mm) Observed Predicted Observed Predicted Observed Predicted Observed Predicted
Year Month

1996 October 63.8 4.8 25.6 0.02 0.15 23 65 23 164
November 122.4 27.4 49.5 0.07 0.09 67 196 115 254
December 127.5 70.6 71.2 0.13 0.18 59 131 211 261
1997 January 182.1 129.5 101.4 0.70 0.23 145 126 437 269
February(110)b 81.8 70.4 45.8 0.23 0.07 64 179 231 229
March(170.7)b 0.0 0.0 0.0 0.00 0.00 0 0 0 0
April 86.5 30.9 26.3 0.15 0.04 115 123 262 169
May 152.4 82.7 70.8 1.10 0.57 123 124 768 388
June 130.3 37.6 31.4 1.24 0.33 117 124 925 291
July 41.1 4.1 3.1 0.12 0.02 9 0 47 0
August(58)b 49.1 0.0 5.7 0.00 0.00 0 0 0 0
September(76)b 0.0 0.0 0.0 0.00 0.00 0 0 0 0
October 85.6 5.5 21.2 0.05 0.19 5 127 37 162
November 56.4 13.1 16.6 0.06 0.29 19 63 124 106
December 133.3 56.8 73.9 0.72 0.37 41 178 314 265
January(142)b
1998 106.6 59.3 69.6 0.58 0.51 39 121 378 174
February(98)b 90.0 36.5 35.3 0.47 0.22 41 55 389 69
March(95)b 88.7 37.7 18.9 0.18 0.08 9 119 86 216
April 130.8 72.6 48.9 0.46 0.43 101 177 468 305
May 111.5 84.6 64.3 0.81 2.08 13 63 748 524
June 31.0 12.3 7.8 0.29 0.09 27 61 144 86




233
(Continued)

© 2007 by Taylor & Francis Group, LLC
234
TABLE 9.4 (CONTINUED)
Monthly Observed Rainfall, Observed and Predicted Runoff, Sediment Loss, Dissolved P, and Total P Losses

Runoff (mm) Sediment Loss (mg/ha) Dissolved P Loss (g/ha) Total P Loss (g/ha)
Rainfalla
(mm) Observed Predicted Observed Predicted Observed Predicted Observed Predicted
Year Month

July 166.1 53.6 48.8 0.23 0.42 142 242 255 463
August(29)b 0.0 0.0 0.0 0.00 0.00 0 0 0 0
September(74)b 0.0 0.0 0.0 0.00 0.00 0 0 0 0
October 27.2 0.0 0.0 0.00 0.00 0 0 0 0
November 141.2 39.9 50.8 0.11 0.70 262 129 394 573
December 205.2 155.0 134.4 0.51 1.51 454 258 1092 1331




Modeling Phosphorus in the Environment
Total (C)c 1992 2661 7446 6299

= 0.52X + 62.3 R2 = 0.45 = 0.71X + 38.5 R2 = 0.61
Regression (C) c,d Y Y

Nash-Sutcliffe 0.35 0.58
Coefficient (C)c

1999 January 224.3 214.8 147.3 1.68 1.89 187 250 1532 1601
February 50.0 7.2 8.1 0.04 0.04 6 56 46 73
March 120.4 58.1 45.9 0.24 0.22 45 122 288 270
April 110.0 65.4 47.5 0.19 0.30 313 122 569 314
May 73.7 6.5 7.0 0.10 0.12 56 51 132 70
June 29.8 0.0 2.2 0.00 0.00 0 0 0 0
July 7.1 0.0 0.1 0.05 0.01 0 0 0 0
August 0.0 0.0 0.0 0.00 0.00 0 0 0 0
September 40.5 0.0 3.6 0.00 0.00 0 0 0 0


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling
October 24.4 0.0 0.7 0.00 0.00 0 0 0 0
November 96.0 1.0 2.7 0.03 0.01 18 65 30 93
December 64.0 11.6 8.9 0.08 0.01 49 123 154 137
January 52.3 9.2 3.4 0.10 0.01 26 0 116 0
2000
February 47.2 7.8 1.3 0.07 0.01 25 0 95 0
March 156.0 77.0 50.5 0.12 0.08 306 127 457 204
April 289.3 213.6 210.0 0.37 0.35 1073 242 1152 592
May 34.8 0.0 0.0 0.00 0.00 0 0 0 0
June 103.6 28.8 26.6 0.28 0.04 99 125 316 162
July 23.9 0.0 0.0 0.00 0.00 0 0 0 0
August 2.0 0.0 0.0 0.00 0.00 0 0 0 0
September 50.5 0.0 0.0 0.00 0.00 0 0 0 0
October 9.4 0.0 0.0 0.00 0.00 0 0 0 0

Total (V)e 4019.8 1785.9 1587.1 11.58 11.66 2203 1283 4888 3516

= 0.82X + 2.63 R = 0.94 = 0.89X + 0.03 R = 0.51 = 0.24X + 34.2 R = 0.51 = 0.83X – 24.1 R2 = 0.88
d,e 2 2 2
Regression (V) Y Y Y Y

Nash-Sutcliffe 0.91 0.22 0.34 0.85
Coefficient (V)e

Notes: Validation period for runoff and sediment was 49 months because no calibration was performed for runoff and sediment simulation. Thus, total, regression
and Nash-Sutcliffe Coefficient under runoff and sediment are for the whole simulation period.
a Rainfall reported under rainfall column reflects only the amount of rainfall associated with monitored data.
b Months when not all storms were successfully monitored for runoff and sediment. The number in parentheses shows total rainfall during that month.
c First 27 months of calibration for P.

d Y = observed; X = predicted.

e Last 22 months of validation for P.




235
© 2007 by Taylor & Francis Group, LLC
236 Modeling Phosphorus in the Environment


organic or inorganic initial soil P content increased both dissolved and sediment-
attached P losses. Based on the sensitivity analysis, attached P loss is more sensitive
to the initial soil organic P than soil inorganic P. Thus, attempts were made to increase
total P loss by increasing the initial soil organic P content, which resulted in an
increase in the dissolved P loss. Because dissolved P loss is more sensitive to the
initial soil inorganic P than soil organic P, attempts were also made to decrease the
dissolved P loss by decreasing the initial soil inorganic P content, which resulted in
a decrease in the total P loss. For final calibration simulation, 750 mg/kg of initial
soil organic P and 25 mg/kg of initial soil inorganic P were used. Because of the
complex and contradictory response of changing initial soil P contents, the regression
of monthly predicted dissolved P loss with observed dissolved P loss resulted in an
R2 of 0.45, slope of 0.52, and E of 0.35. Regression of monthly predicted total P
loss with observed total P loss resulted in an R2 of 0.61, slope of 0.71, and E of
0.58 (Table 9.4).
For the validation study, the results of runoff and sediment simulation (Yuan et al.
2001) are also presented in Table 9.4 because P movement depends on runoff and
sediment movement. No calibration was performed for runoff and sediment simulation
because satisfactory runoff and sediment prediction results were achieved using uncal-
ibrated reference values. Over the 49-month simulation period, AnnAGNPS-predicted
runoff was 89% of the observed total runoff, and AnnAGNPS-predicted dissolved P
loss was 94% of the observed total dissolved P loss (Table 9.4). However, AnnAGNPS-
predicted dissolved P loss was only 58% of the observed dissolved P loss (Table 9.4)
for the validation period. AnnAGNPS-predicted monthly runoff matched well with
observed monthly runoff, but the predicted monthly dissolved P loss did not match
well with the observed monthly dissolved P loss. The regression of the monthly
predicted dissolved P loss with the observed dissolved P loss was fair with an R2 of
0.51, slope of 0.24, and E of 0.34, which also indicated a fair model performance in
predicting dissolved P loss. A time series comparison of observed and predicted
dissolved P loss at the study site (Figure 9.4) shows both overpredictions and under-
predictions by AnnAGNPS.
AnnAGNPS-predicted total P loss during the validation period was 72% of the
observed total P loss (Table 9.4). E was 0.85, which indicated that the model’s
performance in predicting monthly total P loss was good. The regression of the
monthly predicted total P loss with the observed total P loss resulted in an R2 of 0.88
and a slope of 0.83. Although a time-series comparison of observed and predicted
total P loss at the study site (Figure 9.5) shows both overpredictions and underpre-
dictions by AnnAGNPS, observed and predicted total P loss generally coincided better
than observed and predicted dissolved P loss (Figure 9.4 and Figure 9.5).
Both runoff and sediment predictions impact the total P loss prediction. Ann-
AGNPS-predicted sediment loss over the 49-month simulation period was 101% of
the observed sediment loss (Table 9.4). Sediment-attached P loss during the valida-
tion period was 83% of the observed sediment-attached P loss. The observed sedi-
ment-attached P loss was calculated as the difference between the total P loss and
the dissolved P loss.
Simulation results may be improved through a better determination of input
parameters. For example, actual field analysis of crop information at harvest may


© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 237


1200
Observed
Dissolved P loss (g/ha)
1000
Predicted
800

600

400

200

0
Oct-96 Mar-97 Sep-97 Mar-98 Sep-98 Mar-99 Sep-99 Mar-00 Sep-00
Time (months)

FIGURE 9.4 Time series comparison of observed and predicted dissolved P loss.


1800
1600 Observed
Total P loss (g/ha)




1400
Predicted
1200
1000
800
600
400
200
0
Oct-96 Mar-97 Sep-97 Mar-98 Sep-98 Mar-99 Sep-99 Mar-00 Sep-00
Time (months)

FIGURE 9.5 Time series comparison of observed and predicted total P loss.


provide a better estimation of plant P uptake parameters than using literature values.
In addition, after calibration of initial soil P contents, additional calibration of plant
P uptakes may also improve simulation results. Plant P uptake directly impacts the
prediction of dissolved P loss because the crop utilizes dissolved P for growth.
Variation of the solution P would impact the P loss to runoff and sediment because
the amount of P in the solution pool would impact the transfer among three inorganic
pools. Furthermore, uncertainties in soil variables such as content of CaCO3, OC,
clay, base saturation, and soil pH value, which were obtained from the SSURGO
database, would impact the accuracy of the P availability index calculation. The flow
rate between P solution and active pools is determined based on soil moisture,
temperature, P availability index, and the amount of P in each pool. As more soluble


© 2007 by Taylor & Francis Group, LLC
238 Modeling Phosphorus in the Environment


P moves to the active P pool, less soluble P is available for runoff loss. In many
circumstances, the determination of soil- and crop-related input parameters is the
most difficult task for watershed simulations.

9.6 MODEL LIMITATIONS
The difference in the agreement of P losses may be attributed to limitations in Ann
AGNPS simulation of P processes, such as P movement between organic and inorganic
pools and movement of inorganic P between solution and active pools and between
active and stable pools. Such processes are very complicated and difficult to describe
mathematically. AnnAGNPS does not simulate dissolution processes or immobiliza-
tion and leaching of P. Therefore, it is not appropriate to apply Ann-AGNPS to
situations where leaching is an important component. Since the P component simulated
in AnnAGNPS is a simplification of P processes in nature, the model predictions may
be improved with enhancements within AnnAGNPS that more completely describe
the P processes and movements. In addition, enhancement of P in stream processes is
also needed if the model is to be applied to large watersheds.

9.7 CONCLUSIONS
AnnAGNPS P component was described and evaluated in this chapter. AnnAGNPS
models dissolved P in runoff and sediment-attached P resulting from sheet and rill
erosion. To simulate P losses, AnnAGNPS maintains daily soil mass balances of P
for each computational area (AnnAGNPS cell) and keeps track of both organic and
inorganic P mass balances. Major processes considered are residue decomposition
and mineralization, fertilizer application, plant uptake, runoff, and erosion losses.
Application of AnnAGNPS in the DH watershed demonstrated that AnnAGNPS
adequately simulated monthly total P. The simulation of monthly dissolved P was
not as satisfactory as monthly total P. The differences between simulated and
observed results may be attributed to the simplification of P processes in AnnAGNPS
and uncertainties in input selections. Initial soil P contents are the most sensitive
parameters of those selected for sensitivity analysis in determining P losses.


REFERENCES
Baethgen, W.E. and M.M. Alley. 1989. Optimizing soil and fertilizer nitrogen use by inten-
sively managed winter wheat. I: crop nitrogen uptake. Agron. J. 81(1):116–120.
Baginska, B., W. Milne-Home, and P.S. Cornish. 2003. Modeling nutrient transport in Currency
Creek, NSW with AnnAGNPS and PEST. Environ. Model. Softw. 18(8):801–808.
Bagnold, R.A. 1966. An approach to the sediment transport problem from general physics.
Professional Paper 422-J, U.S. Geological Survey, Reston, VA.
Barrow, N.J. and T.C. Shaw. 1975. The slow reactions between soil and anions. 2: effect of
time and temperature on the decrease in phosphate concentration in soil solution. Soil
Sci. 119(2):167–177.
Bassett, D.M., W.D. Anderson, and C.H. Werkhoven. 1970. Dry matter production and nutrient
uptake in irrigated cotton (Gossypium hirsutum). Agron. J. 62(2):299–302.




© 2007 by Taylor & Francis Group, LLC
Phosphorus Modeling 239


Bingner, R.L. and F.D. Theurer. 2001. AnnAGNPS: estimating sediment yield by particle size
for sheet and rill erosion. Proc. 7th Interagency Sediment. Conf. 1:1–7.
Bingner, R.L., F.D. Theurer, and Y. Yuan. 2003. AnnAGNPS technical processes. Available
at http://www.ars.usda.gov/Research/docs.htm?docid=5199.
Boquct, D.J. and G.A. Breitenbeck. 2000. Nitrogen rate effect on partitioning of nitrogen and
dry matter by cotton. Crop Sci. 40(6):1685–1693.
Cronshey, R.G. and F.D. Theurer. 1998. AnnAGNPS — non-point pollutant loading model.
Proc. First Fed. Interagency Hydrol. Model. Conf. 1:9–16.
Fang, F., P.L. Brezonik, D.J. Mulla, and L.K. Hatch. 2002. Estimating runoff phosphorous
losses from calcareous soils in the Minnesota River Basin. J. Environ. Qual.
31(6):1918–1929.
Flannery, R.L. 1986. Plant food uptake in a maximum yield soybean study. Potash and
Phosphate Institute, Conway, AR.
Freeze R.A. and J.A. Cherry. 1979. Groundwater. Englewood Cliffs, NJ: Prentice Hall.
Geter, W.F. and F.D. Theurer. 1998. AnnAGNPS — RUSLE sheet and rill eosion. Proc. First
Fed. Interagency Hydrol. Model. Conf. 1:17–24.
Havlin, J.L., J.D. Beaton, S.L. Tisdale, and W.L. Nelson. 1999. Soil Fertility and Fertilizers:
An Introduction to Nutrient Management, 6th ed. Upper Saddle River, NJ: Prentice
Hall.
Jenson, M.E., R.D. Burman, and R.G. Allen. 1990. Evapotranspiration and irrigation water
requirements. Am. Soc. Civ. Eng. Man. Rep. Eng. Pract. 70. Reston, VA: ASCE.
Johnson, G.L., C. Daly, G.H. Taylor, and C.L. Hanson. 2000. Spatial variability and interpolation
of stochastic weather simulation model parameters. J. Appl. Meteorol. 39(6):778–796.
Jones, C.A., C.V. Cole, A.N. Sharpley, and J.R. Williams. 1984. A simplified soil and plant
phosphorus model. I: documentation. Soil Sci. Soc. Am. J. 48(4):800–805.
Lane, L.J. and V.A. Ferreira. 1980. Sensitivity analysis. Pp. 113–158 in CREAMS: A Field-
Scale Model for Chemicals, Runoff, and Erosion from Agricultural Management
Systems, W.G. Knisel (Ed.). Washington, D.C.: USDA–SEA.
Leonard, R., W.G. Knisel, and D.A. Still. 1987. GLEAMS: Groundwater Loading Effects of
Agricultural Management Systems. Trans. ASAE 30(5):1403.
Merkel, W.A., D. Woodword, and C.D. Clarke. 1988. Ephemeral gully erosion model. Pp.
315–323 in Modeling of Agricultural, Forest, and Range Hydrology, K.L. Campbell
(Ed.). St. Joseph, MI: American Society of Agricultural Engineers.
Mullins, G.L. and C.H. Burmester, 1990. Dry matter, nitrogen, phosphorous accumulation by
four cotton varieties. Agron. J. 82(4):729–736.
Munns, D.N. and R.L. Fox. 1976. The slow reaction which continues after phosphate adsorp-
tion: kinetics and equilibrium in some tropical soils. Soil Sci. Soc. Am. J. 40(1):46–51.
Nash, J.E. and J.V. Sutcliffe. 1970. River flow forecasting through conceptual models. Part
I: a discussion of principles. J. Hydrol. 10(3):282–290.
Nearing, M. A., L. Deer-Ascough, and L. M. Laflen. 1990. Sensitivity analysis of the WEPP
hillslope profile erosion model. Trans. ASAE 33(3): 839–849.
Natural Resources Conservation Service. 2005. Soil Survey Geographic (SSURGO) database,
U.S. Department of Agriculture, Washington, D.C. Available at http://www.ncgc.
nrcs.usda.gov/products/datasets/ssurgo/index.html.
Penman, H.L. 1948. Natural evaporation from open water, bare soil, and grass. Proc. Royal
Soc. London A 193:120–145.
Pote, D.H., T.C. Daniel, D.J. Nichols, A.N. Sharpley, P.A. Moore, Jr., D.M. Miller, and D.R.
Edwards. 1999. Relationship between phosphorous levels in three ultisols and phos-
phorous concentrations in runoff. J. Environ. Qual. 28(1):170–175.




© 2007 by Taylor & Francis Group, LLC
240 Modeling Phosphorus in the Environment


Pote, D.H., T.C. Daniel, A.N. Sharpley, P.A. Moore, Jr., D.R. Edwards, and D.J. Nichols.
1996. Relating extractable soil phosphorous to phosphorous losses in runoff. Soil Sci.
Soc. Am. J. 60(3):855–859.
Rajan, S.S.S. and R.L. Fox. 1972. Phosphate adsorption by soils. 1: influence of time and ionic
environment on phosphate adsorption. Commun. Soil. Sci. Plant Anal. 3:493–504.
Rebich, R.A. 2004. Suspended sediment and agrochemicals in runoff from agricultural sys-
tems in the Mississippi Delta: 1996–2000. Pp. 104–118 in Water Quality Assessment
in the Mississippi Delta, Regional Solutions, National Scope, M.T. Nett, M.A. Locke,
and D.A. Pennington (Eds.). Washington, D.C: ACS.
Renard, K.G., G.R. Foster, G.A. Weesies, D.K. McCool, and D.C. Yoder, coordinators. 1997.
Predicting Soil Erosion by Water: A Guide to Conservation Planning with the Revised
Universal Soil Loss Equation (RUSLE). Washington, D.C.: U.S. Department of Agri-
culture, Agriculture Handbook 703.
Self, J.R. and P.N. Soltanpour. 2004. Soil sampling, Colorado State University Cooperative
Extension. Available at http://www.ext.colostate.edu/pubs/crops/00500.html.
Sharpley, A.N. 1982. Prediction of the water extractable phosphorus content of soil following
a phosphorus addition. J. Environ. Qual. 11(2):166–170.
Sharpley, A.N. 1995. Dependence of runoff phosphorous on extractable soil phosphorous
level. J. Environ. Qual. 24(5):920–926.
Sharpley, A.N., C.A. Jones, C. Gray, and C.V. Cole. 1984. A simplified soil and plant
phosphorus model. II: prediction of labile, organic, and sorbed phosphorus. Soil Sci.
Soc. Am. J. 48(4):805–809.
Sharpley A.N. and J.R. Williams. 1990. EPIC—Erosion Productivity Impact Calculator. 1:
model documentation, Washington, D.C.: U.S. Department of Agriculture, Technical
Bulletin 1768.
Smedema L.K. and D.W. Rycroft. 1983. Land Drainage. Ithaca, NY: Cornell University Press.
Soil Conservation Service. 1985. Section 4: Hydrology in National Engineering Handbook,
Washington, D.C.: U.S. Department of Agriculture.
Suttles, J.B., G. Vellidis, D. Bosch, R. Lowrance, J.M. Sheridan, and E.L. Usery. 2003.
Watershed-scale simulation of sediment and nutrient loads in Georgia coastal plain
streams using the Annualized AGNPS model. Trans. ASAE 46(5):1325–1335.
Theurer, F.D. and C.D. Clarke. 1991. Wash load component for sediment yield modeling.
Proc. 5th Fed. Interagency Sediment Conf. 7:1–8.
Theurer, F.D. and R.G. Cronshey. 1998. AnnAGNPS-reach routing processes. Presented at
the First Federal Interagency Hydrologic Modeling Conference, April 19–23, Las
Vegas, NV.
Young, R.A., C.A. Onstad, D.D. Bosch, and W.P. Anderson. 1989. AGNPS: a nonpoint-source
pollution model for evaluating agricultural watersheds. J. Soil Water Conserv.
44(2):168–173.
Young, R.A., C.A. Onstad, D.D. Bosch, and W.P. Anderson. 1995. AGNPS: an agricultural
nonpoint source model. Pp. 1011–1020 in Computer Models of Watershed Hydrology,
V.P. Singh (Ed.). Highlands Ranch, CO: Water Resources Publications.
Yuan, Y., R.L. Bingner, and R.A. Rebich. 2001. Evaluation of AnnAGNPS on Mississippi
Delta MSEA watersheds. Trans. ASAE 44(5):1183–1190.
Yuan, Y., S.M.M. Debney, and R.L. Bingner. 2002. Cost effectiveness of agricultural BMPs
for sediment reduction in the Mississippi Delta. J. Soil Water Conserv. 57(5):259–267.
Yuan, Y., R.L. Bingner, and R.A. Rebich. 2003. Evaluation of AnnAGNPS nitrogen loading
in an agricultural watershed. J. Am. Water Resour. Assoc. 39(2):457–466.




© 2007 by Taylor & Francis Group, LLC
Đề thi vào lớp 10 môn Toán |  Đáp án đề thi tốt nghiệp |  Đề thi Đại học |  Đề thi thử đại học môn Hóa |  Mẫu đơn xin việc |  Bài tiểu luận mẫu |  Ôn thi cao học 2014 |  Nghiên cứu khoa học |  Lập kế hoạch kinh doanh |  Bảng cân đối kế toán |  Đề thi chứng chỉ Tin học |  Tư tưởng Hồ Chí Minh |  Đề thi chứng chỉ Tiếng anh
Theo dõi chúng tôi
Đồng bộ tài khoản