See discussions, stats, and author profiles for this publication at: [Link]

net/publication/286956313

Evaluating the Transport and Fate of Nutrients in Large Scale River Basins
Using an Integrated Modeling System

Article · September 2013

DOI: 10.1007/978-94-007-6530-6_10

CITATIONS READS

4 971

2 authors, including:

May Wu
Argonne National Laboratory
74 PUBLICATIONS 3,148 CITATIONS

SEE PROFILE

All content following this page was uploaded by May Wu on 06 September 2016.

The user has requested enhancement of the downloaded file.

Chapter 10
Evaluating the Transport and Fate
of Nutrients in Large Scale River Basins
Using an Integrated Modeling System

Zhonglong Zhang and May Wu

Abstract Watershed management is essential in minimizing riverine and

receiving water pollution. No watershed model is currently available that ade-
quately model the large river basin system over long periods of time in a satisfied
level of detail. Watershed models must be linked to riverine or other detailed
receiving water models in order to adequately represent the intricacies of the
physical system under study. Watershed models are used to provide ‘‘boundary
conditions,’’ both hydrologic/hydraulic and nonpoint source loading fluxes, to the
receiving water models. The objectives of this study were to develop an integrated
watershed and riverine modeling system using Soil and Water Assessment Tool
(SWAT) model and Hydrologic Engineering Center-River Analysis System (HEC-
RAS) model to evaluate the transport and fate of nutrients and water quality
impacts in large scale river basins. The SWAT model was constructed for the
Upper Mississippi River Basin (UMRB) in an attempt to account for key elements
associated with crop production and land use changes. The model was calibrated
and validated by using 18 years of observed United States Geological Survey
(USGS) streamflow discharge and water quality data. The SWAT model was used
to predict flow and nutrient exports from each tributary within the watershed. The
results were used as HEC-RAS model’s inputs through an interface. The HEC-
RAS model was able to simulate the transport and fate of nutrients and dynamic
changes in riverine nutrient concentrations. The integrated SWAT and HEC-RAS
modeling system provides a systematic approach to modeling nonpoint nutrient
sources, transport, and fate in a large scale river basin. The modeling system can

Z. Zhang (&)
Badger Technical Services, U.S. Army Engineer Research
and Development Center, 3909 Halls Ferry Road,
Vicksburg, MS 39180, USA
e-mail: [Link]@[Link]
M. Wu
Center for Transportation Research, Energy System Division,
Argonne National Laboratory, 9700 S. Cass Avenue,
Lemont, IL 60439, USA

B. Fu and K. B. Jones (eds.), Landscape Ecology for Sustainable 187

Environment and Culture, DOI: 10.1007/978-94-007-6530-6_10,
Ó Springer Science+Business Media Dordrecht 2013
188 Z. Zhang and M. Wu

be used to predict downstream water quality impacts with land use changes and
assess the effectiveness of different watershed management practices whether
directed toward nutrient source supply or abatement.

Keywords SWAT HEC-RAS Watershed model Riverine model Hydrology





Land use change Sediment Nutrient transport

10.1 Introduction

Nutrient pollution is a leading cause of water quality impairment in lakes and

estuaries and is also a significant issue in rivers (USEPA 2007). The Mississippi
River and tributary streams have been greatly impacted by excess nitrogen,
phosphorus, and sediment loadings from cropland and other sources. The impor-
tance of linkages between watershed processes, and the environmental and eco-
logical responses of aquatic resources are best exemplified by the formation of an
oxygen deprived area in Gulf of Mexico. The hypoxic zone in the Gulf varies in
size from year to year, in 2007, it was the third-largest hypoxic zone on record
(Devine et al. 2008). The hypoxic zone has become a serious threat to commercial
fishing, shrimping, and recreation industries. A number of factors contribute to the
size of the dead zone, but nutrient pollution substantially drives the problem. The
formation of hypoxic zone in the Gulf is a problem caused in large part by
excessive inputs of sediments and nutrients arising from industrial and agricultural
activities in the Mississippi River Basin (MRB) (Diaz and Rosenberg 2008). Soil
erosion that moves sediments and sediment-bound nutrients and pesticides into
waterways is another factor influencing water quality in the Gulf. In particular,
nitrogen is more abundant in dissolved forms, whereas phosphorus is largely
present in particulate forms (either adsorbed or as a constituent of inorganic and
organic particles). As a result, there is a strong correlation between suspended
sediment and total phosphorus concentrations, and changes to the river system that
alter the flow of water or sediment in the system are likely to cause a larger change
in the concentration and transport of phosphorus than of nitrogen (Wetzel 2001).
The evidence now suggests that both nitrogen and phosphorus affect the size of the
dead zone (Devine et al. 2008).
Nutrients enter the Mississippi River and the Gulf from a variety of sources,
including fertilizer runoff from farms, golf courses, and lawns; manure disposal;
discharge from sewage treatment plants and industrial facilities; nitrogen deposi-
tion from the atmosphere; and erosion of nutrient-rich soil. The data provided by
USGS shows that areas with significant agricultural uses are the largest contrib-
utors of nutrient pollution to the Gulf. The United States Geological Survey
(USGS) found that on average, from 2001 to 2005, the Upper Mississippi and
Ohio-Tennessee River basins represent about 31 % of the total land area within the
MRB, they contribute about 82 % of the nitrate-nitrogen flux, 69 % of the total
Kjedahl Nitrogen (TKN), and 58 % of the total phosphorus flux.
10 Evaluating the Transport and Fate of Nutrients 189

The amount of sediments and nutrients transported to the Gulf is not simply a
direct function of what is coming off the field but must also include what’s lost
along the way as water moves through the drainage network (Alexander et al.
2000; Mueller and Spahr 2006). Studies conducted only at fields or small water-
sheds are not able to capture how sedimentation and nitrogen and phosphorus
concentration at multiple scales are influenced by various agricultural practices.
Effective management of water quality and ecosystem, whether directed toward
nutrient supply or abatement, requires both watershed and riverine models to
quantify the transport and fate of nutrients throughout the basin system.
The Soil and Water Assessment Tool (SWAT) model has been widely used to
quantify the flow, sediment and nutrient loadings in large river basins (Gassman
et al. 2007). The SWAT model can be used to identify the location and magnitude
of sediment and nutrient runoff hotspots associated with crop production and land
use changes. The flow-routing methods presently used in SWAT are hydrological
routing methods that are based on the continuity equation and on empirical rela-
tionships to replace the momentum equation. In SWAT, one-dimensional (1D)
hydrologic model is used to mathematically represent flow routing along a river
reach. In this case, simplified schemes, such as linear reservoirs, Muskingum-
Cunge methods may be applied. When dealing with large scale rivers, however,
backwater effect and floodplain inundation may become governing factors for
flood wave routing, and a 1D-hydraulic model is a more suitable method. The
weakness of SWAT models for dynamic flow routing in the large scale watershed
areas is well known. It is very important to perform the flow routing process
accurately because routed results affect other aspects such as sediment routing and
the in-stream nutrient process, both of which are strongly tied to water routing. To
address this problem, a 1D unsteady state flow model (UNET) developed for the
main stem of the Illinois River was coupled with the HSPF model to perform the
flow routing (Lian et al. 2007). SWAT was also coupled with 2D hydrodynamic
and water quality model (CE-QUAL-W2) in the Cedar Creek Reservoir study
(Debele et al. 2008). The results indicated that the two models are compatible and
can be used to assess and manage water resources in complex watersheds com-
prised of watershed and receiving waterbodies.
In this study, an integrated watershed and riverine modeling system was
developed. The water quality impacts of crop production and land use changes in
the Upper Mississippi River Basin (UMRB) were evaluated. The UMRB is the
main focus for future renewable bioenergy sources and provides a majority of
conventional and potential cellulosic feedstock for biofuel production (USDOE
2011). Results of the basin’s SWAT model were used as the mainstem Hydrologic
Engineering Center-River Analysis System (HEC-RAS) model boundary condi-
tions and inputs for evaluating the transport and fate of sediments and nutrients in
the Upper Mississippi River. An integrated modeling system can be used to pre-
dict: (1) Spatial and temporal patterns of nutrient export in the watershed; (2)
Nutrient cycling and microbial and chemical reactions within the river; (3) Long-
term changes in riverine nutrient concentrations and their causes; (4) Effects of
watershed nutrient loading on the downstream water quality; and (5) Effectiveness
190 Z. Zhang and M. Wu

of watershed management scenarios implemented to address problems in the

downstream.

10.2 Upper Mississippi River Basin

The Upper Mississippi River Basin (UMRB) is located in the Midwest United
States and is one of the seven major basins contributing to the Mississippi River
(Fig. 10.1). The Mississippi River is the seventh largest in the world, based on
discharge (580 km3 year-1); whereas the MRB is the third largest in the world
based on area (3,225,000 km2). The Mississippi river flows some 3,770 km
downstream from Lake Itasca, Minnesota, through ten states and eventually dis-
charges to the Gulf of Mexico. The UMRB includes more than 2,011 km of the
Mississippi River extends from the Mississippi River head water at Lake Itasca in
Minnesota to the confluence of Mississippi and Ohio Rivers at Cairo in Illinois.
The Mississippi River is divided into the upper and lower basin at Cairo, Illinois
where the Missouri River enters as the last major tributary. UMRB has the greatest
amount of artificially drained soil, the highest percentage of total land in agri-
culture (corn and soybean) and the highest use of nitrogen fertilizers in the nation.
The Upper Mississippi River (UMR) has been defined by the U.S. Congress as
active commercial navigation routes within the UMRB. The UMRB has a total
drainage area of approximately 490,000 km2, about 15 % of the entire MRB,
including large parts of the states of Illinois, Iowa, Minnesota, Missouri, and
Wisconsin. Small portions of Indiana and South Dakota are also within the basin.

Fig. 10.1 Geographic locations of Mississippi River sub-basins

10 Evaluating the Transport and Fate of Nutrients 191

Soil type in the UMRB ranges from heavy, poorly drained clay soil to light, well-
drained sands. Mollisols dominate but there are also Alfisols that have high clay
subhorizons and lower amounts of organic matter and clay in the surface horizons.
These soils are highly productive but have poor internal drainage. Subsurface tile
drainage of soils in this region is necessary for crop production. Corn and soybean
require a well-drained warm soil for optimum growth. In most parts of the basin,
agriculture is the dominant land use (Fig. 10.2). This figure indicates that over 60 %
of total land is used for agriculture and pasture. Corn, soybeans, and alfalfa are the
major crops in the basin. The UMRB encompasses most of the Corn Belt region of
the U.S. Nitrogen and phosphorus are abundant in the UMRB because of the
widespread use of commercial and animal-manure fertilizers. In fact, the quantity

Fig. 10.2 UMRB soil types

192 Z. Zhang and M. Wu

Fig. 10.3 UMRB land use and land cover

of nitrogen and phosphorous lost from the land to the stream in the UMRB is higher
than in any other portion of the MRB (Fig. 10.3).
The climate of the UMRB is subhumid continental. The climate of the UMRB,
which encompasses most of Illinois, Iowa, Minnesota, Missouri, and Wisconsin, is
humid continental, with warm, moist summers and cold, dry winters. Average
monthly temperatures vary significantly throughout the year, with maximums in
July and minimums in January. The majority of the precipitation (approximately
three-fourths) occurs between April and October. Average annual rainfall for the
area is 91 cm. Monthly average temperatures range from -12 to 0 °C for January
to 22–27 °C in July, with a year-round average of about 11 °C. Flows comprised
of runoff from the UMRB’s stream network support navigation and hydroelectric
plants and fulfill municipal, industrial, and agricultural water requirements.
10 Evaluating the Transport and Fate of Nutrients 193

10.3 Integrated Watershed and Riverine Modeling System

10.3.1 SWAT Model

The SWAT is a lumped parameter watershed model developed and maintained by

the U.S. Department of Agriculture (USDA) Agriculture Research Service (ARS).
SWAT uses algorithms from a number of previous ARS models including
CREAMS, GLEAMS, EPIC, and SWRRB (Arnold et al. 1998). The SWAT model
was developed to assess the impact of land management and climate patterns on
water, sediment, and agricultural chemical yields over long time periods in large
watersheds. The watershed is partitioned into a number of subbasins. Each subbasin
possesses a geographic position in the watershed and is spatially related to adjacent
subbasins. Each subbasin is further divided into hydrological response units (HRU)
based on topography, land use, and soil. HRUs are the smallest computational units
in SWAT with unique land use, soil type and slope within a subbasin. Thus, SWAT
can take two levels of the spatial heterogeneity into account. The first level (sub-
basin) supports the spatial heterogeneity associated with hydrology, and the second
level (HRU) incorporates the spatial heterogeneity associated with land use, soil
type and slope class. Within a subbasin, SWAT does not retain the spatial location
of each HRU. The loss of spatial information within the subbasin introduces a
measure of unrealism and requires caution in interpreting model results. In SWAT,
hydrologic, soil, water quality and other processes are modeled within the subbasins
through the use of HRUs. Flow generation, sediment yield, and pollutant loadings
are summed across all HRUs in a subbasin, and the resulting flow and loads are then
routed through channels, ponds, and/or reservoirs to the watershed outlet. SWAT
typically produces daily results for every subbasin outlet, each of which can be
summed to provide monthly and annual load estimates.
Major model components include climate, hydrology, nutrient cycle, pesticide,
plant growth, and land management. For climate, SWAT uses the data from the
station nearest to the centroid of each subbasin. The hydrology module simulates
major hydrologic components and their interactions as simple responses using
empirical relationships (Fig. 10.4). Precipitation in the form of either rainfall or
snowfall is the major driving mechanism of the hydrologic cycle. SWAT calcu-
lates actual ET based on potential evapotranspiration (PET) from soils and plants
separately. PET can be estimated by three methods: Priestley-Taylor (Priestley and
Taylor 1972), Hargreaves (Hargreaves and Samani 1985), and Penman-Monteith
(Allen et al. 1989). Surface runoff volume and infiltration are computed with the
modified curve number method or Green and Ampt equation. The peak rate
component uses Manning’s formula to determine the watershed time of concen-
tration and considers both overland and channel flow. The soil profile is subdivided
into multiple layers that support soil water processes including infiltration, evap-
oration, plant uptake, lateral flow, and percolation to lower layer. A storage routing
technique is used to calculate redistribution of water between layers in the soil
profile. Lateral subsurface flow in the soil profile is calculated simultaneously with
194 Z. Zhang and M. Wu

Fig. 10.4 SWAT hydrologic processes and interactions

percolation. Groundwater flow contribution to total streamflow is simulated by

routing a shallow aquifer storage component to the stream (Arnold and Allen
1996). Channel routing is simulated using either the variable-storage method or the
Muskingum method; both methods being variations of the kinematic wave model.
Erosion and sediment yield are estimated for each HRU with the Modified Uni-
versal Soil Loss Equation (Williams and Berndt 1977). The channel sediment routing
uses a modification of Bagnold’s sediment transport equation (Bagnold 1977) that
estimates the transport concentration capacity as a function of velocity. The model
either deposits excess sediment or re-entrains sediment through channel erosion
depending on the sediment load entering the channel. The delivery ratio is estimated
for each particle size as a linear function of fall velocity, travel time, and flow depth.
SWAT simulates the complete soil nutrient cycle for nitrogen and phosphorus.
The soil nitrogen cycle is simulated using five different pools; two are inorganic
forms (ammonium and nitrate) while the other three are organic forms (fresh,
stable, and active). Similarly, SWAT simulates six different pools of phosphorus in
soil; three are inorganic forms and the rest are organic forms. Primary biochemical
transformations of nitrogen and phosphorus are simulated. Nitrate export with
runoff, lateral flow, and percolation are estimated as products of the volume of
water and the average concentration of nitrate in the soil layer. Organic nitrogen
and organic phosphorus transport with sediment is calculated with a loading
function developed by McElroy et al. (1976) and modified by Williams and Hann
(1978) for application to individual runoff events. The loading function estimates
daily organic nitrogen and phosphorus runoff loss based on the concentrations of
constituents in the top soil layer, the sediment yield, and an enrichment ratio. The
amount of soluble phosphorus removed in runoff is predicted using labile phos-
phorus concentration in the top soil layer, the runoff volume and a phosphorus soil
10 Evaluating the Transport and Fate of Nutrients 195

partitioning coefficient. In-stream nutrient dynamics in SWAT are simulated using

the kinetic routines from the QUAL2E in-stream water quality model (Brown and
Barnwell 1987).
SWAT allows detailed agricultural management practices to be simulated,
tracking planting, tillage, and fertilization operations and calculating resultant
plant growth with specific dates or with a heat unit scheduling approach during the
year. The plant growth component of SWAT utilizes routines for phenological
plant development based on plant-specific input parameters such as energy and
biomass conversion, temperature, water and nutrient constraints, canopy height
and root depth, and shape of the growth curve. A single plant growth module is
used in SWAT for simulating all crops and assessing removal of water and
nutrients from the root zone, transpiration, and biomass/yield production.

10.3.2 HEC-RAS Model

Few models provide the ability to couple river flow quantity with sediment and
water quality, and those that do are proprietary products that are both expensive as
well as difficult, if not impossible, to modify to suit local conditions. HEC-RAS is
a public domain model developed by the U.S. Army Corp of Engineers (USACE)
([Link] and is widely used and accepted by the engi-
neering community and many regulatory agencies. The HEC-RAS model contains
1D river analysis components for: (1) hydraulic simulation; (2) movable boundary
sediment transport simulation; and (3) water quality analysis. A key element is that
all three components use a common geometric data representation and common
geometric and hydraulic computation routines. In addition to the three river
analysis components, the model contains several hydraulic design features that can
be invoked once the basic water surface profiles are computed, data storage and
management capabilities, graphics and reporting facilities.
The hydraulic simulation is the key computation engine. It performs 1D steady
and unsteady flow calculations on a network of natural or manmade open channels.
Hydraulic calculations are performed at each cross section to compute water
surface elevation, critical depth, energy grade elevation, and velocities. HEC-RAS
is able to perform mixed flow regime (subcritical, supercritical, hydraulic jumps,
and draw downs) calculations in the unsteady flow computations module. The
hydraulic calculations for cross-sections, bridges, culverts, and other hydraulic
structures that were developed for the steady flow component were incorporated
into the unsteady flow module. The model can handle a full network of channels, a
dendritic system, or a single river reach.
Sediment simulation in HEC-RAS utilizes one dimensional, cross-section
averaged, hydraulic properties from RAS’s hydraulic engines to compute sediment
transport rates and update the channel geometry based on sediment continuity
calculations. The sediment module runs in the quasi-unsteady mode, i.e. it com-
putes the unsteady hydraulics as a series of steady state events. The sediment
196 Z. Zhang and M. Wu

transport in HEC-RAS is based on shear stresses computed from these steady state
events. However, the interaction of the bed profile with the entrainment and
transport equations is quasi-dynamic in that the bed is adjusted for erosion or
deposition by the Exner equation. Sediment transport module simulates suspended
load (fine sediments moving at the same speed as water) and bedload (sand and
gravel moving at a slower rate along the bed). Seven different transport functions
are currently available in RAS including Ackers and White, Englund-Hansen,
Laursen, Myer-Peter-Muller, Toffaleti, Yang, and Wilcock. Currently HECRAS
employs Exner 5, a ‘‘three layer’’ algorithm to compute bed sorting mechanisms.
Exner 5 divides the active layer into two sublayers, simulating bed coarsening by
removing fines initially from a thin cover layer. During each time step, the
composition of this cover layer is evaluated and if, according to a rough empirical
relationship, the bed is partially or fully armored, the amount of material available
to satisfy excess capacity can be limited.
HEC-RAS nutrient water quality module includes a set of nutrient simulation
modules (NSM) (Zhang and Johnson 2012). NSM I computes riverine algal bio-
mass, organic and inorganic nitrogen and phosphorus species, CBOD and DO.
NSM II computes multiple algal biomass, nitrogen, phosphorus, and carbon
cycling, DO, COD, alkalinity, pH and pathogen, as well as numerous additional

Fig. 10.5 Schematic representation of HEC-RAS-NSM I

10 Evaluating the Transport and Fate of Nutrients 197

composite constituents. In addition, NSM III incorporates a dynamic bed sediment

diagenesis component, which simulates the chemical and biological processes
undergone at the sediment–water interface after sediments are deposited. The
schematic representation of HEC-RAS NSM I water quality processes is shown
Fig. 10.5. The conservation-of-mass equation in HEC-RAS is solved using
QUICKEST-ULTIMATE transport algorithm.

10.3.3 Integration of SWAT and HEC-RAS

Although SWAT includes riverine water hydraulic and water quality capabilities,
the demands of tracking the dynamic transport and fate of nutrients and/or the
specific characteristics of the receiving water system required additional capabil-
ities not available within SWAT in order to address real-world water management
concerns. Figure 10.6 shows a framework of integrated SWAT and HEC-RAS
modeling system. Evaluation of the transport and fate of nutrients in the watershed
was broken into the following steps:
1. Simulation of the 8 digit HUCs was conducted using a SWAT model
2. Extraction of time series SWAT outputs of runoff and water quality from each HUC
watershed along the mainstem and used as boundary conditions to HEC-RAS.
3. Simulation of the Mainstem River Using a HEC-RAS Model
4. Assess the potential impacts of nutrient loading and land use changes in the basin.
As shown in Fig. 10.6, SWAT is linked to the HEC-RAS model providing
tributary discharge and contaminant loads to the HEC-RAS model. Once the
SWAT model was satisfactorily developed and calibrated the SWAT model results
are imported into the Hydraulic Engineering Center’s Data Storage System (HEC-
DSS) through an interface (Fig. 10.7). For the HEC-RAS model, SWAT provides
discharge and contaminant loads from the major streams and drainages tributary to

Fig. 10.6 Integated SWAT

and HEC-RAS modeling
system
198 Z. Zhang and M. Wu

Fig. 10.7 HEC-DSS importer interface from SWAT outputs

the HEC-RAS model’s domain as its boundary conditions. The HEC-RAS model
is then calibrated based on measured in-stream flow, sediment and nutrient con-
centrations. The goal is not only to characterize each of those features but
sequentially to link them together as part of a systematic evaluation of the large
river basin based on a defined set of conditions.
Table 10.1 describes the state variables that SWAT passes to the HEC-RAS
model. The state variables required for HEC-RAS are passed from SWAT using
HEC-DSS. As shown in Table 10.1, SWAT provides daily time-series for flow and
contaminant loads for HEC-RAS.

Table 10.1 SWAT state variables passed to HEC-RAS

SWAT source HEC-RAS input Unit
Flow Flow cms
Suspended sediment Suspended sediment ton/day
NH4-N NH4-N kg/day
NO3-N NO3-N kg/day
NO2-N NO2-N kg/day
Organic N Organic N kg/day
PO4-P PO4-P kg/day
Organic P Organic P kg/day
Algae Algae kg/day
DO DO kg/day
CBOD CBOD kg/day
10 Evaluating the Transport and Fate of Nutrients 199

10.4 Application of the Integrated Modeling System

to the Upper Mississippi River Basin

The SWAT model inputs consist of topography, soil properties, land use/cover
type, weather/climate data, and land management practices. The study watershed
is divided into subbasins. The USGS divided the UMRB into 131 eight-digit-level
subbasins with an average drainage area of 3,755 km2 (Fig. 10.8). Each subbasin
is further divided into hydrological response units (HRU) based on topography,
land use, and soil.
We have calibrated and validated the SWAT outputs against observed data.
Those results showed that the SWAT model properly reproduced observed flow,
sediment, total nitrogen and phosphorus from the basin. We now have used those
calibrated outputs from the SWAT model as input into the HEC-RAS model and
calibrate the hydraulic and water quality processes in the UMR.
The dominant cause of nutrient flux to the UMR is agricultural activities in the
UMRB. Corn and soybean fields are the largest source of N input (52 %) followed
by atmospheric deposition (16 %). Among agricultural land uses, the largest P
input was from pasture and range lands (animal manure; 37 %) followed by corn
and soybeans (25 %). In contrast to N, urban sources (12 %) made a significant
contribution of P to the UMR (Alexander et al. 2000). The SWAT model reveals a
consistent trend of heavy nitrogen and phosphorus exports from areas of intensive
agricultural production in the UMRB.
The SWAT model, in addition to runoff, outputs sediments and water quality
variables, such as total suspended solids (TSS), nitrogen species (organic nitrogen,
nitrate/nitrite, and ammonia/ammonium), phosphorus species (organic and mineral
phosphorus), chlorophyll-a, dissolved oxygen (DO), and carbonaceous biochemical
oxygen demand (CBOD). The HEC-RAS/NSM I model has the same water quality
state variables as the SWAT model and accepts inputs in terms of one-to-one
relationship. We used a separate interface tool that imports SWAT outputs into an
appropriate time-series dataset within the HEC-DSS for HEC-RAS. The interface
extracts daily SWAT outputs (runoff and its water quality constituents) at required
reach or subbasin outlets into HEC-DSS file that is acceptable to HEC-RAS.
Principal tributaries of the UMR are the Minnesota, St. Croix, Wisconsin, Rock,
Iowa, Des Moines, Illinois, and Missouri Rivers and several smaller rivers and
streams. In the 1,076 km of river between the first lock, Upper St. Anthony Falls,
and the last lock of the Channel Navigation Project, Lock 27, the UMR falls 128 m
with an average slope of approximately 9.5 cm/km. Average flow of the UMR
ranges from 280 m3 s-1 at St. Paul, Minnesota, to 4,955 m3 s-1 at St. Louis,
Missouri. Data input for HEC-RAS included geometric data to represent river
networks, channel cross-section data, and hydraulic structure data such as bridges
and culvert data. Cross-section data includes station-elevation data, main channel
bank stations, downstream reach lengths, roughness coefficients, and contraction
and expansion coefficients. Hydrologic events are represented by flow data. Time-
series predicted by SWAT for all of the required constituents and drainage sources
200 Z. Zhang and M. Wu

Fig. 10.8 SWAT UMRB subbasin distributions

to the mainstem river will provide the necessary boundary conditions to exercise
HEC-RAS.
Modeled flow discharge, suspended sediment, total nitrogen and total phos-
phorus fluxes for a number of UMRB long-term sampling sites between 1990
through 2009 were compared with observed data. Figures 10.9, 10.10, 10.11 and
10.12 show the monthly modeled and observed comparison for the simulation
period. Discharge and water quality data from the UMR were acquired from the
US Geological Survey (USGS).
The difference between the measured and modeled monthly average stream
flow for all four gages was less than 10 % for the simulation period. The hydro-
graph plot revealed a strong correlation between the measured and modeled flows
as indicated by the R2 and NSE values, which were 0.76 and 0.72, respectively at
10 Evaluating the Transport and Fate of Nutrients 201

Fig. 10.9 Observed versus modeled monthly streamflow for the simulation period

Fig. 10.10 Observed versus modeled monthly TSS for the simulation period

the watershed outlet. Compared with the streamflow results, the simulations on
suspended sediment loads were less accurate. The modeled monthly TN and TP
loads generally followed the month-to-month observed trends. The seasonal var-
iation of the monthly streamflow TN and TP loads was fairly well reproduced in
the calibration period. These results show a reasonable match between the mod-
eled results and observed data, indicating the model’s ability to adequately sim-
ulate both the watershed and in-stream processes involved within the UMRB.
Watershed characteristics, including tributary streams, point and non-point
pollution sources, all influence mainstem river water quality. Variations in land use
practices, cover types, and watershed area will determine the level and type of
sediment, nutrient, and contaminant inputs into the UMR from their tributaries.
Typical seasonal fluctuations in flow, as well as periodic extreme events, can have
dramatic effects on river water quality. This modeling system is able to quantify
the effects of agricultural production systems on nitrogen and phosphorus runoff to
streams and the nutrient cycling in the waters delivered to the watershed oulet.
202 Z. Zhang and M. Wu

Fig. 10.11 Observed versus modeled monthly TN for the simulation period

Fig. 10.12 Observed versus modeled monthly TP for the simulation period

Efficient reductions may also be achieved by targeting both nitrogen and phos-
phorus sources in close proximity to large rivers.

10.5 Summary

The purpose of this study was to develop an integrated watershed and riverine
modeling system and use this integrated modeling system to study the transport
and fate of nutrients in a large scale river basin. The SWAT model and the HEC-
RAS model were loosely integrated so that output from the watershed model
became input into the riverine model. The model integration considers spatial and
temporal characteristics of the watershed systems, correspondence and transfer-
ence of the state variables between two models, and file format specifics for proper
communications between the models.
10 Evaluating the Transport and Fate of Nutrients 203

The SWAT model was successfully used to estimate flow and nutrient exports
from each tributary in the UMRB. The HEC-RAS model was able to use daily
SWAT outputs and model nutrient cycling and transport and fate of nutrients in the
mainstem of the basin. Land use changes and different watershed management
scenarios could be evaluated using the integrated modeling system to determine
water quality impacts. The integrated watershed and riverine modeling system
provides a systematic approach to evaluating nonpoint nutrient sources, transport,
and fate in a large river basin. This approach will provide an effective management
tool to investigate changes in the flow quantity and water quality in the Mississippi
River.

References

Alexander RB, Smith RA, Schwarz GE. Effect of stream channel size on the delivery of nitrogen
to the Gulf of Mexico. Nature. 2000;403:758–61.
Allen RG, Jensen ME, Wright JL, Burman RD. Operational estimates of evapotranspiration.
Agron J. 1989;81:650–62.
Arnold JG, Allen PM. Estimating hydrologic budgets for three Illinois watersheds. J Hydrol.
1996;176:55–77.
Arnold JG, Srinivasan R, Muttiah RS, Williams JR. Large-area hydrologic modeling and
assessment: part I. Model Dev J AWRA. 1998;34(1):73–89.
Bagnold RA. Bed-load transport by natural rivers. Water Resour Res. 1977;13:303–12.
Brown CL, Barnwell Jr. To the enhanced stream water quality models QUAL2E and QUAL2E-
UNCAS documentation and user manual: USEPA. Environmental research laboratory.
Athens. Georgia. 1987. EPA/600/3-85/040. p. 455.
Debele B, Srinivasan R, Parlange J-Y. Coupling upland watershed and downstream waterbody
hydrodynamic and water quality models (SWAT and CE-QUAL-W2) for better water
resources management in complex river basins. Environ Model Assess. 2008;13:135–53.
Devine J, Dorfman M, Rosselot KS. Missing Protection: Polluting the Mississippi river basin’s
small streams and wetlands. October: Natural resources defense council issue paper; 2008.
Diaz RJ, Rosenberg R. Spreading dead zones and consequences for marine ecosystems. Science.
2008;321:926–8.
Gassman PW, Reyes MR, Green CH, Arnold JG. The soil and water assessment tool: historical
development, applications, and future research directions. Trans ASABE. 2007;50(4):1211–50.
Hargreaves GH, Samani ZA. Reference crop evapotranspiration from temperature. Appl Eng
Agric. 1985;1:96–9.
Lian Y, Chan I-C, Singh J, Demissie M, Knapp V, Xie H. Coupling of hydrologic and hydraulic
models for the Illinois river Basin. J Hydrol. 2007;344:210–22.
McElroy AD, Chiu SY, Nebgen JW, Aleti A, Bennett FW. Loading functions for assessment of
water pollution from nonpoint sources. Environ Prot Tech Serv.1976; EPA 600/2-76-151.
Mueller DK, Spahr NE. Nutrients in streams and rivers across the Nation—1992–2001: U.S.
geological survey scientific investigations report 2006–5107 (2006).
Priestley CHB, Taylor RJ. On the assessment of surface heat flux and evaporation using large-
scale parameters. Mon Weather Rev. 1972;100:81–2.
USDOE (U.S. Department of Energy) U.S. Billion Ton Update: Biomass supply for a bioenergy
and bioproducts industry. Perlack RD and Stokes BJ (Leads). ORNL/TM-2011/224. Oak
Ridge National Laboratory, Oak Ridge, TN.2011, p. 227.
USEPA (U.S. Environmental Protection Agency). Hypoxia in the Northern Gulf of Mexico: an
update by the EPA Science Advisory Board, 2007. EPA-SAB-08-003.
 204 Z. Zhang and M. Wu

Wetzel RG. Limnology, lake and river ecosystems. 3rd ed. San Diego: Academic Press; 2001.
p. 1006.
Williams JR, Berndt HD. Sediment yield prediction based on watershed hydrology. Trans ASAE.
1977;1100–04.
Williams JR, Hann RW. Optimal operation of large agricultural watersheds with water quality
constraints. Texas Water Resources Institute, Texas A&M University, Technical Report. No. 96,
1978
Zhang Z, Johnson BE. Aquatic nutrient simulation modules (NSM). ERDC/EL-12-X, Draft, U.S.
Army Engineer Research and Development Center, Vicksburg, MS. 2012.

View publication stats