Content uploaded by Mahesh Tapas
Author content
All content in this area was uploaded by Mahesh Tapas on Sep 28, 2024
Content may be subject to copyright.
INTEGRATIVE ANALYSIS OF POLICY CHANGES FOR A COASTAL WATERSHED:
IMPLICATIONS FOR AGRICULTURE AND ECOSYSTEM HEALTH
By
Mahesh Ramesh Tapas
July, 2024
Director of Dissertation: Randall Etheridge, PhD
Major Department: Coastal Studies
ABSTRACT
This dissertation explores the integration of hydrological modeling, the effects of sea level rise,
and socio-economic factors influencing watershed management in North Carolina, primarily
utilizing the Soil and Water Assessment Tool Plus (SWAT+) with four independent publication
chapters. The first chapter assesses the effectiveness of various satellite precipitation products and
autocalibration techniques on river flow prediction, highlighting the superior performance of the
Global Precipitation Measurement Integrated Multi-satellitE Retrievals (GPM IMERG) dataset
when combined with the Generalized Likelihood Uncertainty Estimation (GLUE) technique. The
second study investigates the effects of sea level rise on nitrate dynamics within the Tar-Pamlico
coastal watershed, with adjustments made to SWAT+ parameters to simulate changes in nitrogen
processes and their impacts on ecosystem health. This reveals increased nitrate loads under sea
level rise scenarios. The third chapter merges econometric and engineering frameworks to evaluate
the efficacy of agricultural best management practices (BMPs) as influenced by farmers'
behavioral responses. It reveals that despite potential incentives, significant reductions in nitrate
loading are not achieved, underscoring the limitations of current models and the importance of
comprehensive socio-hydrological frameworks. Collectively, this dissertation enhances our
understanding of hydrological processes and their interactions with environmental changes and
human factors, offering crucial insights for effective watershed management and policy
development.
Integrative Analysis of Policy Changes for a Coastal Watershed: Implications for Agriculture
and Ecosystem Health
A Dissertation
Presented to the Faculty of the Department of Coastal Studies
East Carolina University
In Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Integrated Coastal Sciences
By
Mahesh Ramesh Tapas
July, 2024
Director of Dissertation: Randall Etheridge, PhD
Dissertation Committee Members:
Ariane Peralta, PhD
Gregory Howard, PhD
Natasha Bell, PhD
Venkataraman Lakshmi, PhD
© Mahesh Ramesh Tapas, 2024
DEDICATION
I dedicate this dissertation to my family, who have been my greatest support. My mother, Sangita
Tapas, has always been there for me with her love and strength. My sister, Ankita Tapas, has been
a constant source of support and joy. And to my late father, Ramesh Tapas, I am forever grateful
for his wisdom and guidance. Their support has been invaluable, and this achievement is because
of their belief in me.
ACKNOWLEDGMENTS
I would like to express my sincere gratitude to everyone who has supported me throughout my
PhD journey. A heartfelt thank you to my dissertation committee members—Dr. Randall
Etheridge, Dr. Gregory Howard, Dr. Ariane Peralta, Dr. Natasha Bell, and Dr. Venkataraman
Lakshmi—for their invaluable guidance and support. A special acknowledgment to Dr. Randall
Etheridge, who has been the best advisor in the world, providing unparalleled mentorship and
encouragement. I am also grateful to the SWAT+ developers and my collaborators for their
contributions and teamwork that have enriched my research. Additionally, I want to thank my
fellow PhD students and the faculty in the ICS PhD cohort for their encouragement. I also
acknowledge the generous funding from the National Science Foundation and the Environmental
Protection Agency, which was crucial to my research. Your collective support has been crucial to
my success.
TABLE OF CONTENTS
List of Tables ...................................................................................................................................x
List of Figures ............................................................................................................................... xii
CHAPTER 1: SATELLITE-BASED RAINFALL DATASETS AND AUTOCALIBRATION
TECHNIQUES’ EFFECTS ON SWAT+ FLOW SIMULATION ..................................................1
Highlights .............................................................................................................................2
Abstract ................................................................................................................................2
Keywords .............................................................................................................................3
Graphical Abstract ...............................................................................................................3
Introduction ..........................................................................................................................4
Materials and Methods .........................................................................................................8
Study Area ...............................................................................................................8
SWAT+ model .......................................................................................................10
Datasets ..................................................................................................................11
Model set up ............................................................................................11
Description of SPPs ...................................................................................12
ERA-5 ............................................................................................12
gridMET .........................................................................................12
IMERG ...........................................................................................13
In-situ data .................................................................................................14
Autocalibration techniques ....................................................................................14
Performance metrics ..............................................................................................17
R-SWAT ................................................................................................................18
Results ................................................................................................................................21
Sensitivity analysis.................................................................................................21
Analysis of simulated streamflow ..........................................................................24
SWAT+ calibration results evaluation ...................................................................28
Parameter uncertainty analysis ..............................................................................30
Discussion ..........................................................................................................................32
Sensitivity analysis under different rainfall datasets .............................................32
Flow simulation, rainfall dataset, and optimization technique ..............................34
Effects of rainfall dataset and optimization technique on performance indices ....37
Conclusion .........................................................................................................................39
Acknowledgment ...............................................................................................................41
Data availability .................................................................................................................41
References ..........................................................................................................................42
CHAPTER 2: SIMULATING THE EFFECTS OF SEA LEVEL RISE ON NITROGEN
EXPORT USING THE SWAT+ MODEL: A CASE STUDY OF THE TAR-PAMLICO RIVER
BASIN, NC, USA ..........................................................................................................................60
Abstract ..............................................................................................................................60
Study region .......................................................................................................................61
Keywords ...........................................................................................................................61
Highlights ...........................................................................................................................62
Graphical abstract ..............................................................................................................62
Introduction ........................................................................................................................63
Objectives ..............................................................................................................66
Materials and methods .......................................................................................................67
Study area...............................................................................................................67
SWAT+ model setup .............................................................................................68
SWAT+ model optimization ..................................................................................70
SLR incorporation ..................................................................................................71
Results and discussion .......................................................................................................74
Model optimization ................................................................................................74
Modeling the effects of SLR on nitrogen processes ..............................................75
Simulated effects on crop yields ................................................................79
Simulated effects on denitrification rate ....................................................80
Nitrate load to the estuary ......................................................................................81
Conclusion .........................................................................................................................85
References ..........................................................................................................................86
CHAPTER 3A: FRAMEWORK FOR STAKEHOLDER-DRIVEN SOCIO-HYDROLOGICAL
MODELING: CONCEPTUAL FOUNDATIONS FOR POLICY DEVELOPMENT AND
EVALUATION TO IMPROVE ECOSYSTEM HEALTH ........................................................100
Abstract ............................................................................................................................100
Keywords .........................................................................................................................101
Introduction ......................................................................................................................101
Context and challenge ..........................................................................................101
Objective ..............................................................................................................102
Significance..........................................................................................................102
Theoretical background of socio-hydrological integration ..............................................103
Socio-hydrology ...................................................................................................103
Challenges of integration .....................................................................................104
Econometric modeling in socio-hydrological contexts ....................................................104
Basics ...................................................................................................................104
Advantages ...........................................................................................................105
Framework for integrating engineering and economic models .......................................106
Methodological overview ....................................................................................106
Potential applications ...........................................................................................110
Broader implications ............................................................................................111
Conclusion and future directions .....................................................................................111
References ........................................................................................................................112
CHAPTER 3B: A SOCIO-HYDROLOGICAL FRAMEWORK FOR INCORPORATING
FARMERS’ BEHAVIORAL MODEL INTO THE SWAT+ MODEL AND IMPLICATIONS
FOR IMPROVING ECOSYSTEM HEALTH ...........................................................................123
Abstract ............................................................................................................................123
Study area.........................................................................................................................124
Keywords .........................................................................................................................124
Highlights .........................................................................................................................125
Graphical abstract ............................................................................................................125
Introduction ......................................................................................................................126
Study objectives ...................................................................................................129
Materials and methodology..............................................................................................129
Study area.............................................................................................................130
Hydrological model: engineering framework ......................................................130
SWAT+ model calibration .......................................................................132
Best management practices simulation ....................................................136
Farmers’ behavioral model: Econometrics framework ........................................136
Choice experiment design in farmer preference survey ..........................137
Estimation of Willingness to Accept (WTA) ..........................................141
Hydrological-Econometric models integration ....................................................143
Results and discussion .....................................................................................................145
SWAT+ model optimization ................................................................................145
Farmers behavioral modeling ..............................................................................146
Effects of reduced N-fertilizer application rate and cover crops on watershed
processes and N-load ...........................................................................................150
Effects of 30% reduced fertilizer application on nitrate load ..................150
Effects of Cover Crops on Nitrate Load ..................................................154
Hydrological and Socio-Hydrological Model Comparison .................................159
Area Conversion and Budget Allocation .................................................163
Agricultural Output and Environmental Impact ......................................164
Economic Efficiency ................................................................................165
Conclusions ......................................................................................................................167
References ........................................................................................................................170
LIST OF TABLES
Table 1.1: Characteristics of the Tar-Pamlico and Cape Fear watersheds ....................................10
Table 1.2: Data used for HRU creation .........................................................................................12
Table 1.3: Description of SPP data sets .........................................................................................13
Table 1.4: SWAT+ flow parameters selected for sensitivity analysis and optimization ...............16
Table 1.5: Model performance indicators used in this study ( is observed streamflow,
is simulated streamflow, i is ith simulation, and is the mean value, and n is the total
number of values) .........................................................................................................................18
Table 1.6: Type of change and range of parameters used for calibration ......................................20
Table 2.1: parameters altered simulate plant yield and denitrification under SLR .......................77
Table 2.2: Average seasonal nitrate load from Tar-Pamlico River Basin .....................................82
Table 2.3 Statistical summary of paired t-test analysis for nitrate loads under baseline and SLR
conditions .......................................................................................................................................84
Table 3b.1: SWAT+ model optimization parameters details [bsn: Basin, sol: Soil, hru:
Hydrological Response Unit, plt: Plant] .....................................................................................134
Table 3b.2: Policy scenarios for hydrological and socio-hydrological modeling .......................144
Table 3b.3: Mixed Logit Model Results ......................................................................................147
Table 3b.4: Seasonal average flow and nitrate load (Kg N) with reduced fertilizer application rate
......................................................................................................................................................152
Table 3b.5: Reduced fertilizer application ANOVA (single factor) test at Pamlico estuary .......155
Table 3b.6: Seasonal variation of flow and nitrate load (kg N) with cover crops .......................156
Table 3b.7: Summary of ANOVA results for cover crop scenarios ............................................157
Table 3b.8 Cost-benefit analysis of different engineering and econometric scenarios ...............161
LIST OF FIGURES
Figure 1.1 The location of (A) Cape Fear and (B) Tar-Pamlico watersheds ...................................9
Figure 1.2 Parameter sensitivity analysis (a: CF_ERA-5, b: CF_gridMET, c: CF_GPM IMERG,
d: TP_ERA-5, e:TP_gridMET, f:TP_GPM IMERG; * indicates t_stat<20; # indicates t_stat>10)
........................................................................................................................................................23
Figure 1.3 Tar-Pamlico flow simulation variation under different scenarios (a: ERA-5,
b:gridMET, c: GPM IMERG) .......................................................................................................26
Figure 1.4 Cape Fear flow simulation variation under different scenarios (a: ERA-5, b:gridMET,
c: GPM IMERG) ...........................................................................................................................27
Figure 1.5 Performance indices for different scenarios (TP: Tar-Pamlico, CF: Cape Fear) ........30
Figure 1.6 Variation of the common sensitive flow parameters for model combinations and both
watersheds: Tar-Pamlico (TP), Cape Fear (CF) ...........................................................................32
Figure 2.1 (a) Location of the Tar-Pamlico River basin within the United States, (b) Digital
elevation model and geographical characteristics of the Tar-Pamlico watershed, (c) Hydrological
monitoring station in Washington used for model calibration, (d) Tar-Pamlico’s coastal region
where we incorporated the affects of SLR .....................................................................................68
Figure 2.2: Monthly nitrate load optimization ...............................................................................75
Figure 2.3: Simulated effects of SLR on annual average crop yields ............................................79
Figure 2.4: Impacts of SLR on annual average denitrification [AGRR: General agricultural crop;
SOYB: Soybean] ...........................................................................................................................80
Figure 2.5 Simulated Nitrate Load For SLR and Baseline Scenarios ............................................81
Figure 3a.1 Framework for Integrating SWAT+ and Econometric Models in Socio-Hydrological
Studies ..........................................................................................................................................108
Figure 3b.1 Study Area (Tar-Pamlico watershed) ......................................................................130
Figure 3b.2 Sample cost share contract .......................................................................................138
Figure 3b.3 Simulated and Observed Nitrate Load .....................................................................145
Figure 3b.4 Effects of reduced fertilizer on nitrate load for the baseline, 12% (scenario 4), and
24% (scenario 5) of total agricultural area in the watershed .......................................................151
Figure 3b.5 Effects of reduced fertilizer on monthly nitrate load to the Pamlico estuary for the
baseline, 12% (scenario 2), and 24% (scenario 3) coverage with cover crops ...........................155
Chapter 1: Satellite-Based Rainfall Datasets and Autocalibration Techniques’ Effects on
SWAT+ Flow Simulation
Mahesh R Tapas1*,†, Randall Etheridge2,†, Thanh-Nhan-Duc Tran3, Manh-Hung Le4,5, Brian
Hinckley 2, Van Tam Nguyen6, Venkataraman Lakshmi3
1 Integrated Coastal Program, East Carolina University, Greenville, NC 27858, USA
2 Department of Engineering, Center for Sustainable Energy and Environmental Engineering, East
Carolina University, Greenville, NC 27858, USA
3 Department of Civil and Environment Engineering, University of Virginia, Charlottesville, VA
22904, USA
4 NASA Goddard Space Flight Center Hydrological Sciences Laboratory, MD 20771, USA
5 Science Applications International Corporation, Greenbelt, MD 20771, USA
6 Helmholtz Centre for Environmental Research – UFZ, Saxony 04138, Germany
*Corresponding author: Mahesh R Tapas (tapasm21@students.ecu.edu), 3101, Life Sciences
and Biotechnology Building, East Carolina University, Greenville, North Carolina, USA, 27858
†This author contributes equally to this paper.
Chapter 1
2
Highlights
• Cn2, revap_co, flo_min, revap_min, and awc are the most sensitive parameters.
• Flow simulation is influenced more by rainfall forcing than autocalibration method.
• GPM IMERG rainfall with the GLUE technique showed the highest performance.
Abstract
Accurate flow prediction is a primary goal of hydrological modeling studies, which can be affected
by the use of varying rainfall datasets, autocalibration methods, and performance indices. The
combined effect of three satellite precipitation products (SPPs) — Fifth generation of European
ReAnalysis (ERA-5), Gridded meteorological data (gridMET), Global Precipitation Measurement
Integrated Multi-satellitE Retrievals (GPM IMERG) — and three autocalibration techniques —
Dynamically Dimensioned Search (DDS), Generalized Likelihood Uncertainty Estimation
(GLUE), Latin Hypercube Sampling (LHS) — on Soil Water Assessment Tool Plus (SWAT+)
river flow prediction is measured using three evaluation metrics — Nash Sutcliffe Efficiency
(NSE), Kling Gupta Efficiency (KGE) and coefficient of determination (R2) — for two watersheds
in North Carolina (Cape Fear, Tar-Pamlico) using the SWAT+ model. Our findings are: (1) Five
parameters in the SWAT+ model, cn2, revap_co, flo_min, revap_min, and awc are the most
sensitive among others and over chosen watersheds. (2) Rainfall products are found to have a
greater role in streamflow variation compared to calibration techniques. (3) GPM IMERG
outperforms other SPPs over assessments, followed by ERA-5 and gridMET. (4) NSE is more
sensitive to SPPs and calibration techniques than the KGE. (5) SWAT+ underperforms in the
prediction of baseflow for the groundwater-driven watershed. Overall, our findings suggest that
the GPM IMERG rainfall dataset, combined with the GLUE optimization technique and KGE
Chapter 1
3
performance index, provides the most effective approach for optimal flow simulations. The results
from this study will help hydrological modelers choose an optimal combination of rainfall dataset,
autocalibration technique, and performance index depending on watershed characteristics.
Keywords
Hydrological Modeling; SWAT+; Rainfall Datasets; Autocalibration Techniques; Performance
Indices
Graphical Abstract
Chapter 1
4
1. Introduction
Hydrological models are developed and used to support policies at the watershed and regional
scales (Brauman et al., 2022; Aryal et al., 2022). These models are designed using a variety of
mathematical equations (e.g., continuity equation, Darcy’s law, and Manning’s equation) to
simulate the complex interactions of hydrological processes (Grimaldi et al., 2021; Herman et al.,
2020 ; Ahmed et al., 2020). Policymakers use models’ results to evaluate the effectiveness of
proposed policies, improve effectiveness of policies, and mitigate unintended side effects when
implemented (Maviza & Ahmed, 2021; Pandi et al., 2021; Nguyen et al., 2022; Tran et al., 2021a,
2021b, 2022c; Nuong et al., 2022). These models serve as valuable tools for showing how water
quality and quantity may change within a watershed due to alterations in land cover management
practices (Brauman et al., 2022; Tran, 2022d; Smigaj et al., 2023). However, the accuracy of these
multi-parameterized models has been questioned for informed decision-making (Garibaldi et al.,
2019; Grimaldi et al., 2021).
In recent decades, the rapid growth in computational power has enabled the development
of more complex models that use a higher number of parameters to accurately represent water
movement in a watershed (Adeyeri et al., 2020). However, there are often limited numbers of
observations available to assist with model calibration, raising challenges for the accuracy of
model predictions (Gupta & Govindaraju, 2019). This mismatch between observed data and multi-
parameter simulations increases modeling uncertainty (Beven & Freer, 2001; Wellen et al., 2015).
Another type of uncertainty in environmental system models comes from input datasets (Bosshard
et al., 2013). These datasets include rainfall, land cover, elevation, and soil information (Chaubey
et al., 2005; Cho et al., 2009). Uncertainty in the rainfall dataset comes from low spatiotemporal
Chapter 1
5
resolution and assumptions made to average rainfall over the spatial scale used in the model (Fraga
et al., 2019; Junqueira et al., 2022).
In recent years, many studies have examined the use of satellite-based rainfall data sets as
an alternative to in-situ data, such as Sharifi et al. (2019), Tran et al. (2023b), or Le et al. (2020)
evaluating the performance of common satellite-based precipitation products (SSPs) in Austria
and Vietnam, respectively. Specifically, Tran et al. (2023b) and Le et al. (2020) conducted
experiments that showcased that Global Precipitation Measurement (GPM) Integrated Multi-
satellitE Retrievals (IMERG) (Hou et al., 2014) achieved the highest performance across different
assessments. However, Tran et al. (2023b) mentioned that SPPs tend to underestimate the total
actual amount of rainfall and have issues with rainfall estimates in coastal and arid regions. Thus,
one of the main objectives of this study is to evaluate some commonly used precipitation datasets,
namely the Fifth Generation of European ReAnalysis (ERA-5) (Muñoz-sabater et al., 2021),
IMERG (Hou et al., 2014) and the Gridded Surface Meteorological (gridMET) (Abatzoglou, 2013)
to optimize flow simulations in a coastal region.
The Soil and Water Assessment Tool (SWAT), developed by the United States Department
of Agriculture (USDA) and Texas A&M AgriLife, is a comprehensive hydrological model
designed to predict the impacts of land use, land management, and climate change on sediment
transport (Betrie et al., 2011), water quality (Oeurng et al., 2016), and nutrient cycling (Kansara et
al., 2021) in river basins. This model has been widely used in previous studies throughout the
world: Mekong River Basin (Mohammed et al., 2018b, 2018a; Mondal et al., 2022), Vietnam River
Basins (Le et al., 2020; Tran et al., 2022a, 2022b, 2023c), Nepal (Kumar et al., 2017), and United
States (Jha et al., 2006; Sehgal et al., 2018; Tapas et al., 2022a, 2022b, Tran et al., 2022e). In recent
years, the new version of SWAT, known as SWAT plus (SWAT+), has been released with new
Chapter 1
6
features regarding spatial process interactions and visualization (Bieger et al., 2017; Wu et al.,
2023). This version allows users to better model the transport and retention of nutrients and
sediment in the watershed.
Hydrological models depend on various simplifications, making it extremely challenging
to obtain accurate simulations with the initial model setup (Gupta & Govindaraju, 2019; Vieux,
2001). To address this issue and improve the correlation between simulated and observed flow,
hydrological modelers perform calibration. Although manual calibration by adjusting individual
parameters is possible, the high number of parameter combinations makes this approach
impractical (Efstratiadis & Koutsoyiannis, 2010; Sunmin, 2021). Consequently, extensive study
has been devoted to developing more accurate hydrological models using automatic optimization
techniques (Getirana, 2010; Yen et al., 2019). Arsenault et al. (2014) compared ten stochastic
calibration methods and determined that Dynamically Dimensioned Search (DDS) (Tolson &
Shoemaker, 2007) outperforms most other techniques, especially with increased model
complexity. The Generalized Likelihood Uncertainty Estimation (GLUE) is a widely utilized
method among hydrological modelers due to its simplicity (Blasone et al., 2008). Latin Hypercube
Sampling (LHS) (Stein, 1987) is another popular method in uncertainty analysis, as it prevents
repeated sampling (Yue et al., 2023). However, there is still a lack of comprehensive understanding
regarding the application of these calibration methods when using satellite-based precipitation
products (SPPs) in the context of enhancing flow simulations at the basin scale.
The performance of hydrological models can be assessed using visual interpretations or
statistical indices (Jain & Sudheer, 2008; Wealands et al., 2005). Visual analysis of simulated
streamflow can give useful insights into model behavior (Wealands et al., 2005) but needs expert
analysis, and is impractical from an automatic calibration point of view (Jain & Sudheer, 2008).
Chapter 1
7
However, performance indices do not require expert opinions and can be optimized using
automated calibration techniques (Idrissou et al., 2020; Jain & Sudheer, 2008; Ritter & Munoz-
Carpena, 2013). Several performance indices are available to measure the agreement between the
observed data and simulated data. Nash Sutcliffe Efficiency (NSE; Nash & Sutcliffe, 1970) is the
most widely used performance index for many hydrological models (Tran et al., 2023a, 2023b; Du
et al., 2022); however, it is advisable to assess hydrological models using additional performance
indices along with NSE (Jain & Sudheer, 2008; Ritter & Munoz-Carpena, 2013; Tran et al.,
2023d). Other popular performance indices with hydrological modelers are Kling Gupta Efficiency
(KGE) (Gupta et al., 2009) and coefficient of determination (R2) (Adeyeri et al., 2020; Idrissou et
al., 2020; Jain & Sudheer, 2008; Chang et al., 2023). NSE, KGE, and R2 are popular as they can
be compared with their ideal value of one. All these indices have shortcomings and strengths. For
example, NSE does not distinguish between different kinds of errors (Jain & Sudheer, 2008), KGE
corrects bias factors from NSE (Gupta et al., 2009), and R2 is not able to measure non-linear
relationships among different parameters (Barrett, 1974; Waseem et al., 2017).
The optimal flow prediction and parameter set can vary based on the rainfall dataset,
autocalibration techniques, and evaluation metric used. However, there are few studies that provide
a comprehensive examination of the relationship between these factors. This study aims to
examine the combined effects of these factors on SWAT+ flow simulations. We will perform this
analysis on two major watersheds of North Carolina. We will use three different sources of rainfall
data (ERA-5, gridMET, GPM IMERG), three automatic calibration methods (DDS, LHS, and
GLUE), and three evaluation metrics (NSE, KGE, R2). The objectives of this study are (1) to
explore which combinations of three rainfall datasets and three optimization techniques yield
better results for flow simulations; (2) to identify how performance indices vary based on the
Chapter 1
8
chosen rainfall dataset and optimization technique; and (3) to compare parameter sensitivity and
uncertainty analysis for the two watersheds and three rainfall datasets. Our findings will assist
hydrological modelers in choosing the most suitable combination of rainfall datasets, optimization
techniques, and performance index for stream and river flow simulation.
2. Materials and Methods
2.1. Study area
This study was conducted in (1) the Tar-Pamlico watershed and (2) the Cape Fear watershed in
North Carolina (Figure 1.1). The Cape Fear is the largest river basin in the state (NCDEQ, n.d.-a),
while the Tar-Pamlico is the fourth largest (NCDEQ, n.d.-b). These watersheds are two of only
four river basins located entirely within North Carolina. The Cape Fear basin discharges into the
Atlantic Ocean, while the Tar-Pamlico River flows into the Pamlico Sound. Both watersheds boast
diverse ecosystems with a variety of habitats (NCDEQ, n.d.-b, n.d.-a). The Tar-Pamlico watershed
contains a larger proportion of agricultural land and wetlands, whereas the Cape Fear watershed
has a higher percentage of forested land, pastureland, and urban areas. The Cape Fear watershed
also has a higher elevation range compared to the Tar-Pamlico watershed. The major
characteristics of both watersheds are summarized in Table 1.1.
Chapter 1
9
Figure 1.1 The location of (A) Cape Fear and (B) Tar-Pamlico watersheds.
Chapter 1
10
Table 1.1 Characteristics of the Tar-Pamlico and Cape Fear watersheds (sources: (NCDEQ, n.d.-
b, a, 2023).
Characteristics
Tar-Pamlico
Cape-Fear
Location
North Carolina, USA
North Carolina, USA
Area (mi2)
6,400
9,164
Counties
15
26
Population
470,000+
2,070,000+
Land use distribution (%)
(i) Agriculture
(ii) Forest land
(iii) Wetlands
(iv) Pasture land
(v) Rangeland
(vi) Urban areas
27.9
33.9
31.9
3.5
1.3
1.4
18.9
52
17.9
6.4
0
4.7
2.2. SWAT+ model
SWAT is a semi-distributed hydrological model developed in the early 1990s and is one of the
most widely used tools for water quality and quantity simulation (Arnold et al., 1998; Gassman et
al., 2022; Guo & Su, 2019; Tran et al., 2024). Technological advancements since the original
model was developed have made more detailed hydrological modeling possible, and complexity
has been added to the model (Bieger et al., 2017; Efstratiadis & Koutsoyiannis, 2010; Tumsa et
al., 2022). SWAT+ is an advanced version of SWAT with respect to data management, data
analysis, data visualization, and process interaction at a watershed and sub-watershed level
Chapter 1
11
(SWAT+, 2020). SWAT+ is capable of implementing landscape units as well as flow and pollutant
routing over the landscape (SWAT+, 2020). Although the basic equations for calculating flow in
SWAT+ have not been changed, significant changes have been made in the code and the
formatting of the input files (SWAT+, 2020; White et al., 2022). SWAT+ is also more appealing
for spatial and temporal data analysis and data visualization (Bieger et al., 2017). In this study,
version 2.2.0 of SWAT+ (released Feb 2023) was used.
2.3. Datasets
2.3.1. Model set up
We obtained both watershed boundary shapefiles using the United States Geological survey
(USGS) StreamStats website (Ries et al., 2017). We used QSWAT+ (Dile et al., 2019) to create
streams, sub-basins, and Hydrological Response Units (HRU) using the Digital Elevation Model
(DEM), land cover, and soil files. SWAT+ divides the watershed into sub-watersheds, sub-
watersheds into landscape units, and landscape units into HRUs. An HRU is the smallest spatial
unit in SWAT+, with at least one HRU per sub-basin, incorporating unique land cover, soil, and
slope data for their creation (SWAT+, 2020). We used Digital Elevation Model (DEM) data from
the United States Geological Survey (USGS), land-use land cover data from the National Land
Cover Database (NLCD), and soil data from Soil Survey Geographic Database (SSURGO) at 90
m resolution (Table 1.2) within QSWAT+ (Dile et al., 2019) for initial model setup. We simulated
both watersheds with a two-year warm-up period to initiate and stabilize several hydrological
processes (Oo et al., 2020).
Chapter 1
12
Table 1.2. Data used for HRU creation (TP: Tar-Pamlico watershed, CF: Cape Fear Watershed).
Product
Spatial
resolution
Year
Source
DEM
90 m
2011:TP; 2011: CF
USGS
Land-use land cover
90 m
2008:TP; 2010: CF
NLCD
Soil
90 m
2015
SSURGO
2.3.2. Description of SPPs
2.3.2.1. ERA-5
ERA-5 is the latest climate reanalysis product from the European Centre for Medium-Range
Weather Forecasts (ECMWF) and stands as their fifth-generation reanalysis. It provides hourly
data encompassing atmospheric, land-surface, and sea-state parameters, complete with uncertainty
estimates (C3S, 2017; Hersbach et al., 2020; Muñoz-sabater et al., 2021). The ERA-5 analysis is
produced at hourly intervals using an advanced 4Dvar integration approach. This dataset can be
found in the Climate Data Store (CDS), structured on regular latitude-longitude grids with a
resolution of 0.25° × 0.25°. For our study, we obtained the daily ERA-5 precipitation data by
aggregating it from the hourly dataset. The ERA-5 data used in our study is available from the
ECMWF website (https://cds.climate.copernicus.eu/).
2.3.2.2. gridMET
gridMET is a 0.04° resolution daily meteorological dataset providing climate data such as
temperature, precipitation, wind, humidity, and radiation for the contiguous United States
(CONUS). It combines the spatial characteristics of gridded climate data from the Parameter-
elevation Regressions on Independent Slopes Model (PRISM) with the favorable temporal features
Chapter 1
13
(and additional variables) of the regional reanalysis of North American Land Data Assimilation
System Phase 2 (NLDAS-2) through climatically aided interpolation (Abatzoglou, 2013;
Abatzoglou & Ficklin, 2017). The gridMET data used in this study can be accessed from the
Climatology Lab website (https://www.climatologylab.org/gridmet.html).
2.3.2.3. IMERG
IMERG is a multi-satellite precipitation dataset that provides global precipitation data with high
temporal and spatial resolution. It is a product of NASA’s GPM mission, which aims to provide
accurate and comprehensive precipitation measurements for the entire globe (Li et al., 2021). The
GPM IMERG dataset combines data from various sources, including passive microwave sensors,
infrared sensors, and radar, to provide precipitation estimates with a spatial resolution of 0.1o and
a temporal resolution of 30 minutes (Hou et al., 2014). The IMERG product used in this study can
be obtained from the NASA website (https://gpm.nasa.gov/data).
Table 1.3. Description of SPP data sets.
Product
Temporal
coverage
Spatial
resolution
Spatial
coverage
Temporal
resolution
References
ERA-5
1940 –
present
0.25o
90°S –
90°N
24h
Hersbach et al.
(2020)
gridMET
1979 –
present
0.04o
25°N –
50°N
24h
Abatzoglou.
(2013)
IMERG
2000 – 2021
0.1o
60°N –
60°S
1/2h
Hou et al. (2014)
Chapter 1
14
2.3.3. In-situ data
For calibration of the SWAT+ model, we used daily observed flow data from USGS. We used the
USGS station at the Tar River in Greenville, NC (USGS station id: 02084000) for Tar-Pamlico
watershed calibration. The data available at this station is from Oct 1st, 2001, to the present, with
data missing from Oct 2006 to Sept 2007 and Oct 2016 to Aug 2019. We used the USGS station
at Kelly, NC (USGS station id: 02105769) for calibration of the model for the Cape Fear
watershed. The USGS station in Greenville, NC, has some tidal influence and occasionally shows
negative flow values. It is important to note that only a small fraction of the readings—specifically,
50 out of 5022, which is less than 1%—exhibited negative flows. As SWAT+ cannot handle
backflow (SWAT+, 2020), we converted negative flow values to zeros to ensure the negative flows
are captured as low flows and should simulate as low as possible, ideally zero, considering one-
dimensional flow limitations. The USGS station for the Cape Fear River calibration does not have
any tidal influence. Considering the missing observed flow data and available weather data period,
we simulated the Tar-Pamlico watershed from Jan 2001 to Sept 2016 and the Cape Fear watershed
from Jan 2003 to Dec 2018 with two years as a warm-up period.
2.4. Autocalibration techniques
Calibration is an important step in hydrological modeling in which users try to match simulated
data with observed data by changing multiple parameters in the model (Sunmin, 2021; Tolson &
Shoemaker, 2007). We used LHS (Stein, 1987), DDS (Tolson & Shoemaker, 2007), and GLUE
(Blasone et al., 2008) automatic calibration approaches. LHS is a powerful method of selecting
parameter values for a given number of simulations. It divides n number of parameters in m equal
ranges so that each region is targeted. This method tests the combination of parameters in fewer
simulations relative to random sampling. We performed a sensitivity analysis using LHS with
Chapter 1
15
2,100 simulations considering the 17 most widely used flow parameters (Table 1.4). We selected
diverse parameters from the aquifer, soil, channel, HRU, and basin level to comprehensively model
flow (Bailey et al., 2020; Nguyen et al., 2022; Srinivasan et al., 2010; Tan & Yang, 2020). In our
preliminary analysis using the GPM IMERG rainfall dataset, we conducted 2,100 simulations and
found that the sensitive parameters were consistently identified for both the Tar-Pamlico and Cape
Fear watersheds.
The DDS algorithm is a heuristic-artificial intelligence technique to find the best option
amongst given options – the algorithm searches for a globally optimal solution at the start and
narrows to the optimal local solution towards the end of user-specified iterations (Tolson &
Shoemaker, 2007). For the DDS method, we performed parallel processing with 175 iterations on
12 threads with a total of 2,100 simulations. GLUE is a simple and flexible algorithm widely used
by environmental system modelers (Beven & Binley, 1992; Tolson & Shoemaker, 2007). In the
GLUE method, numerous parameter combinations are randomly selected. Each combination is
assigned a likelihood — representing its chance of appearing in multiple model sets — based on
the principle of equifinality, which acknowledges that multiple parameter sets can similarly well
simulate system behavior despite the differences in the parameter values themselves. (Blasone et
al., 2008; Mirzaei et al., 2015). For GLUE and LHS methods, we also used a parallel processing
approach with a total of 2,100 simulations.
Chapter 1
16
Table 1.4. SWAT+ flow parameters selected for sensitivity analysis and optimization.
Parameter
Resolution
Description
alpha
Aquifer
Baseflow recession factor
awc
Soil
Available water capacity of the soil layer
bf_max
Aquifer
Maximum baseflow rate when the whole area is adding up to
the baseflow
canmx
HRU
The upper limit of canopy storage
chk
Channel
Channel base conductivity
cn2
HRU
SCS curve number for soil moisture condition-2
cn3_swf
HRU
Curve number condition III for soil moisture factor
epco
HRU
Plant uptake compensation factor
esco
HRU
Soil evaporation compensation factor
flo_min
Aquifer
The minimum depth from the surface to the water table
required for groundwater flow to occur
k
Soil
Hydraulic conductivity
ovn
HRU
Manning’s coefficient for overland flow
perco
HRU
Percolation coefficient
revap_co
Aquifer
Groundwater revap coefficient
revap_min
Aquifer
Threshold depth of shallow aquifer for revap to occur
surlag
Basin
The coefficient for surface runoff lag
evlai
Basin
Leaf area index at zero evaporation from water bodies
Chapter 1
17
2.5. Performance metrics
The goodness-of-fit of model output is measured using performance indicators (Jain & Sudheer,
2008; Wealands et al., 2005). We used the three most widely used performance indicators: NSE
(Nash & Sutcliffe, 1970), KGE (Gupta et al., 2009), and R2 (Barrett, 1974; Ozer, 1985). All of
them have an ideal value of one for comparison, and each measures different aspects of goodness-
of-fit (Ritter & Muñoz-Carpena, 2013). NSE indicates the variation between observed and residual
data variance (Nash & Sutcliffe, 1970). KGE is a comprehensive score considering the ratios of
means and dispersion, along with the correlation of observed and simulated datasets (Gupta et al.,
2009). The R2 explains how well the simulated variable explains the observed variable’s variation
(Ozer, 1985). Most hydrological models use NSE for testing model performance; however, NSE
can be biased depending on data repetition, magnitude, and the number of measurements (Jain &
Sudheer, 2008; Ritter & Muñoz-Carpena, 2013). Thus, it is advised to assess the model goodness-
of-fit using multiple performance indices (Jain & Sudheer, 2008). In this study, we compared
model results using three performance indices. We optimized the model for NSE and found the
maximum value from all 2,100 simulations using each autocalibration technique. The formula for
calculating NSE, KGE, and R2 are shown in Table 1.5.
Chapter 1
18
Table 1.5. Model performance indicators used in this study (is observed streamflow, is
simulated streamflow, i is ith simulation, and
is the mean value, and n is the total number of
values).
Performanc
e indicator
Metric equation
NSE
KGE
R2
2.6. R-SWAT
Over the last decade, there has been a significant development in autocalibration tools and
platforms for hydrological models. The current SWAT+ calibration tools are SWAT+ Toolbox
(Celray, 2022; Gassman et al., 2007), SWATplusR (Schuerz, 2022), SWATplus-CUP (WWEE,
2023), R-SWAT (Nguyen et al., 2022), and IPEAT+ (Yen et al., 2019). SWAT+ toolbox is coded
in the C language and is not available for the Linux and Mac operating systems. There is no explicit
SWAT+ toolbox user portal functional for researchers to discuss modeling challenges.
SWATplusR is a package in R that allows users to identify reasonable parameter ranges along with
parallel processing. SWATplusR currently does not have a graphical user interface and could be
Chapter 1
19
challenging to use for those with no experience using R software. SWATplusCUP has a graphical
user interface and can be used by beginners. SWATplusCUP premium features, such as parallel
processing, are not open source. IPEAT+ in the current version does not provide uncertainty
analysis or a graphical user interface (Nguyen et al., 2022; Yen et al., 2019).
R-SWAT has a graphical user interface within R and provides an open-source parallel
processing option for SWAT+. R-SWAT was developed using the Shiny-R-package (Nguyen et
al., 2022) and has a user group for researchers to discuss questions. These advantages are why R-
SWAT was used in this study. Parameters can be altered in 3 ways in SWAT+: (1) absolute value
change, (2) change the value by relative percentage, and (3) change the value by a specified
amount. For instance, if the original parameter value in SWAT+ is x, then the new value of a
parameter (x’) after assigning change (C) under different change types will be absolute value
change (x’ = x + c), percentage change (x’ = x + c*x), replace (x’ = c) (Nguyen et al., 2022). The
calibration (CAL) file included with SWAT+ shows the absolute minimum and maximum range
of the parameters in the model. We used the replace parameter change option for the basin-wide
parameters and a few aquifer-level parameters based on previous literature (Table 1.6). We justify
using replace options for aquifer-level parameters due to SWAT+ defaulting to assigning uniform
values for all aquifers in a given watershed; thus, we have no loss of resolution at the aquifer level.
For parameters with a narrow range (range 0 to 1), we used absolute value change, and for the
parameters with a wide range, we used the relative change option (Table 1.6). This allowed us to
target for large number of parameter combinations with limited simulations.
Chapter 1
20
Table 1.6. Type of change and range of parameters used for calibration.
Parameter
Type of change
Min
Max
Units
alpha.aqu
replace
0.01
0.5
Days
awc.sol
absolute
0.01
0.3
mm_H20/mm
bf_max.aqu
absolute
0
1
mm
canmx.hru
relative
-0.25
0.25
mm/H20
chk.rte
relative
-0.25
0.25
mm/hr
cn2.hru
relative
-0.3
0.2
Null
cn3_swf.hru
absolute
0
0.5
Null
epco.hru
absolute
0
0.5
Null
esco.hru
absolute
0
0.5
Null
flo_min.aqu
relative
-0.25
0.5
m
k.sol
relative
-0.25
0.25
mm/hr
ovn.hru
absolute
0
10
Null
perco.hru
absolute
0
0.5
fraction
revap_co.aqu
absolute
0
0.1
Null
revap_min.aqu
relative
-0.25
0.25
m
surlag.bsn
replace
0.05
24
Days
evlai.bsn
replace
0
10
Null
Both watersheds were automatically calibrated using three different methods (LHS, DDS,
GLUE) with three rainfall datasets (ERA-5, gridMET, GPM IMERG) in R-SWAT. Based on the
literature review and watershed characteristics, we selected the 17 most widely used flow
Chapter 1
21
parameters (Table 1.6) for sensitivity analysis and model optimization. We analyzed models with
three performance indices (NSE, KGE, R2) and visual interpretation to explore the goodness-of-
fit of SWAT+ simulated flow for different scenarios.
Sensitivity analysis was conducted utilizing the Latin Hypercube Sampling (LHS) in
parallel mode through R-SWAT. A total of 2,100 simulations were performed using 17 calibration
parameters (Table 1.6). An α value of 0.05 was used to determine whether a parameter was
considered sensitive - a parameter with a p-value less than 0.05 was indicated as sensitive. These
sensitive parameters were confirmed with a high t-statistic (t-stat) value.
3. Results
3.1. Sensitivity analysis
Key parameters affected by the different rainfall datasets were identified using sensitivity analysis
for the Tar-Pamlico and Cape Fear watersheds. For Tar-Pamlico watershed for (a) ERA-5 rainfall
data we found sensitive parameters are cn2, revap_co, flo_min, revap_min, awc, alpha, perco, and
epco; (b) gridMET dataset the sensitive parameters are cn2, revap_co, flo_min, revap_min, awc,
epco and cn3_swf; (c) GPM IMERG dataset we found sensitive parameters are cn2, revap_co,
flo_min, revap_min, awc, alpha, and esco (Figure 1.2).
In the case of the Cape Fear river basin (a) ERA-5 rainfall data we found sensitive
parameters are cn2, revap_co, flo_min, revap_min, awc, alpha, perco, and epco; (b) gridMET
dataset the sensitive parameters are cn2, revap_co, flo_min, revap_min, awc, alpha, perco, epco
and cn3_swf; (c) GPM IMERG dataset we found sensitive parameters are cn2, revap_co, flo_min,
revap_min, awc, alpha, perco, and cn3_swf (Figure 1.2).
Chapter 1
22
Five common sensitive parameters were identified, including cn2, revap_co, flo_min,
revap_min, and awc, for different rainfall datasets for Cape Fear and Tar-Pamlico watersheds
(Figure 1.2). Flow was sensitive to the alpha parameter (representing baseflow) for all rainfall
datasets and both watersheds, with the exception of the Tar-Pamlico watershed model utilizing the
gridMET rainfall dataset. The percolation coefficient (perco) parameter was found to be sensitive
under all rainfall datasets in the Cape Fear watershed and the ERA-5 dataset for the Tar-Pamlico
watershed. For gridMET, flow simulations for both watersheds were found to be sensitive to epco
(plant uptake compensation factor) and cn3_swf (soil water factor for cn3 parameter). In this work,
the chosen watersheds showed their unique characteristics due to complex interactions between
land and water processes, especially under tidal influences which significantly impact the baseflow
and peaks. Our findings can serve as a reference for using these chosen parameters, tested across
various SPPs and calibration techniques. Additionally, they highlight the challenges and
considerations necessary when modeling coastal watershed dynamics.
Chapter 1
23
Figure 1.2 Parameter sensitivity analysis (a: CF_ERA-5, b: CF_gridMET, c: CF_GPM IMERG,
d: TP_ERA-5, e:TP_gridMET, f:TP_GPM IMERG; * indicates t_stat<20; # indicates t_stat>10).
Chapter 1
24
3.2. Analysis of simulated streamflow
We tested nine different simulated flow scenarios with the combination of three rainfall datasets
and three autocalibration techniques using SWAT+ for each watershed. One method of assessing
model accuracy is visually comparing the simulated and observed flow. As shown in Figure 1.3
for the Tar-Pamlico watershed, the high observed flow in 2011 (above 250 m3s-1) occurred around
late August and early September (Figure 1.3). This increased flow was due to Hurricane Irene - a
category one hurricane - hitting eastern North Carolina. The high flows were measured at USGS
Greenville station from August 28th, 2011, to September 5th, 2011, with the highest observed daily
flow on August 31st, 2011, of 345 m3s-1 (Figure 1.3). We simulated the flow for Tar-Pamlico from
2001 to 2016, but to get more insights into SWAT+ peak flow and baseflow predictions, we have
shown the predictions for the year 2011 in Figure 1.3 for the Tar-Pamlico watershed. This section
helps us understand the combined effects of rainfall datasets and optimization techniques on
simulated storm flows. SWAT+ fell short of capturing peak flow duration for all scenarios by three
days. The calibrations using the GPM IMERG rainfall product did the best job predicting the
magnitude of peak flow, whereas gridMET did the worst. Simulated flows using all three rainfall
products under all optimization techniques underestimated the time to peak flow and subsequent
baseflow recession times.
SWAT+ model performed well in predicting flow magnitude with GPM IMERG rainfall
dataset followed by ERA-5 and gridMET dataset. For GPM IMERG data, the GLUE optimization
technique (358 m3s-1) performed better than LHS (426 m3s-1) and DDS (309 m3s-1) for simulating
the peak flow during Hurricane Irene. For ERA-5 data, LHS (311 m3s-1) and GLUE (372 m3s-1)
optimization procedures performed better than DDS (595 m3s-1), for simulating the peak flow. For
gridMET data, the LHS (243 m3s-1) optimization method performed much better than both GLUE
Chapter 1
25
(974 m3s-1) and DDS (1013 m3s-1) for simulating the peak flow during Hurricane Irene, as DDS
and GLUE provided unreasonably high flows.
Similar data is shown in Figure 1.4 for the Cape Fear basin. The results in the Cape Fear
basin for each calibration method and rainfall dataset were similar to the results in the Tar-Pamlico
basin. Figures 1.3 and 1.4 include periods of baseflow and storm flow simulation using different
rainfall data sets. We found that the combinations of the ERA-5 dataset with LHS or GLUE,
gridMET dataset with LHS, and GPM IMERG dataset with either GLUE or LHS give better
baseflow magnitude prediction results. In the Cape Fear watershed, we observed high baseflows
(greater than 40 m3 s-1). LHS performed better than DDS and GLUE in most cases, implying that
the LHS technique should be used over DDS and GLUE for groundwater-driven watersheds
(Figure 1.4).
Chapter 1
26
Figure 1.3 Tar-Pamlico flow simulation for 2011 under different scenarios (a: ERA-5, b:
gridMET, c: GPM IMERG).
Chapter 1
27
Figure 1.4 Cape Fear flow simulation for 2011 under different scenarios (a: ERA-5, b: gridMET,
c: GPM IMERG).
Chapter 1
28
3.3. SWAT+ calibration results evaluation
In this study, we systematically calibrated flow parameters to maximize the Nash-Sutcliffe
Efficiency (NSE) utilizing three distinct rainfall datasets and three optimization techniques within
two different watersheds. KGE and R2 were evaluated based on all 2,100 simulations for a
watershed with a specific set of rainfall data. Figure 1.5 depicts the maximum value of the
particular performance metric amongst all simulations for the given scenario.
GPM IMERG data produced optimal flow for both watersheds under LHS and GLUE
methods. For the Tar-Pamlico watershed, we found the best KGE (0.424) and R2 (0.501) for the
ERA-5 dataset with the LHS method and the best NSE (0.323) with the GLUE method. The
gridMET dataset gave the best KGE (0.188) and R2 (0.439) with the LHS method and the best
NSE (0.059) with the GLUE method. The GPM IMERG rainfall data produced the best KGE
(0.619) with the LHS method and the best NSE (0.412) and R2 (0.498) with the GLUE method
(Figure 1.5).
For the Cape Fear River basin, we found the best NSE (0.449), KGE (0.522), and R2 (0.559)
for ERA-5 data with the LHS method. The gridMET dataset gave the best NSE (0.341) and R2
(0.498) with the GLUE method and the best KGE (0.426) with the LHS method. For the GPM
IMERG dataset, we found the best NSE (0.498) and KGE (0.676) with the GLUE method and the
best R2 (0.532) with the LHS method (Figure 1.5).
The ERA-5 and GPM IMERG datasets performed better than the gridMET rainfall dataset
when evaluated using KGE and NSE indices (Figure 1.5). On the other hand, R2 performed better
than NSE and KGE for the gridMET dataset (Figure 1.5). We found the worst performance index
values for gridMET rainfall data when optimized with the DDS technique (NSE, KGE, R2 — Tar-
Pamlico: -0.236, 0.057, 0.212 —Cape Fear: 0.029, 0.148, 0.304).
Chapter 1
29
We found that KGE varied less among different scenarios than NSE (Figure 1.5). In most
cases, we found higher performance indices for ERA-5 and GMP IMERG rainfall datasets with
the GLUE optimization technique (Figure 1.5). We found comparatively better results for the
gridMET dataset with the LHS optimization technique (NSE, KGE, R2 — Tar-Pamlico: 0.041,
0.188, 0.439 — Cape Fear: 0.314, 0.426, 0.489). NSE gave lower values than KGE and R2 under
all scenarios (Figure 1.5). Maximum values of KGE and R2 varied based on the rainfall dataset
and optimization technique. We found R2 to give consistently higher values than KGE and NSE
for the gridMET rainfall dataset (Figure 1.5). Overall, our findings suggest that the GPM IMERG
rainfall dataset, combined with the GLUE optimization technique and KGE performance index,
provides the most effective approach for optimal flow simulations (Figure 1.5).
Chapter 1
30
Figure 1.5 Performance indices for different SPP and optimization technique scenarios (TP: Tar-
Pamlico, CF: Cape Fear).
3.4. Parameter uncertainty analysis
We conducted a parameter uncertainty analysis to assess the variability and reliability of the
parameters used in predicting streamflow. This analysis encompassed the nine scenarios, involving
three rainfall datasets and three optimization techniques, applied across both watersheds, as
depicted in Figure 1.6. We compared the optimal values variation of the sensitive parameters under
Chapter 1
31
the best NSE for a given scenario. Figure 1.6 shows the box plot of the five common sensitive
parameters (cn2, revap_co, awc, flo_min, and revap_min) for both watersheds. We used relative
change (cn2, flo_min, revap_min) and absolute value change (awc, revap_co) for these parameters
(Table 1.6).
We assumed that the uncertainty in the prediction of these parameters is only from the
rainfall dataset and optimization technique used, as we kept all other data (soil map, land cover,
observed data, and other SWAT+ input data) constant while building the model. We found the
largest variation in the flo_min (Tar-Pamlico: 0.03 to 0.48; Cape Fear: -0.11 to 0.49) and
revap_min (Tar-Pamlico: -0.25 to 0.21; Cape Fear: -0.22 to 0.18) parameters between both
watersheds, which implies those are more uncertain than the other three parameters. We found
comparatively less variation amongst other sensitive parameters for awc (Tar-Pamlico: 0.01 to
0.29; Cape Fear: 0.01to 0.26), cn2 (Tar-Pamlico: -0.29 to 0; Cape Fear: -0.29 to -0.06), and
revap_co (Tar-Pamlico: 0 to 0.09; Cape Fear: 0 to 0.08).
Chapter 1
32
Figure 1.6 Variation of the common sensitive flow parameters for model combinations and both
watersheds: Tar-Pamlico (TP), Cape Fear (CF).
4. Discussion
4.1. Sensitivity analysis under different rainfall datasets
In the analysis of both Tar-Pamlico and Cape Fear watersheds under each of the three distinct
rainfall datasets, five common sensitive parameters consistently emerged: cn2, revap_co, flo_min,
revap_min, and awc. This uniformity underscores the significant reliance of flow on these factors
that control surface runoff and infiltration partitioning, evapotranspiration, and groundwater
discharge to streams. The consistent appearance of these parameters across different scenarios
emphasizes their critical role in influencing and regulating watershed responses to different rainfall
Chapter 1
33
inputs, reinforcing their importance in SWAT+ modeling and understanding of watershed
biogeochemical processes. Our results align with previous literature where they found these
parameters sensitive for streamflow simulation (Wagner et al., 2022; Reza & Salmani 2023; Yen
et al., 2019)
The high sensitivity of alpha parameter for most scenarios showed the significance of
groundwater flow response to recharge changes in flow simulation for ERA-5 and GPM IMERG
rainfall datasets and its dependency on the watershed under the gridMET rainfall dataset. In
SWAT+, the perco parameter plays a critical role in streamflow simulation, demonstrating
sensitivity across most scenarios. This parameter is derived using the Green-Ampt equation for
infiltration, as detailed by Rawls et al. (1983), which significantly influences the model's
hydrological response. This parameter represents the groundwater infiltration rate as a function of
soil hydraulic properties, such as soil texture, structure, and pore size distribution, as well as soil
water content and other factors that affect the soil’s ability to transmit water (SWAT+, 2020).
The epco parameter specifically addresses simulation of plant water uptake in SWAT+. It
is a dimensionless factor that adjusts the plant’s root water uptake to account for the reduction in
potential evapotranspiration (PET) due to water stress. Sensitivity to the epco parameter can be
anticipated in regions with high evapotranspiration rates, such as the Cape Fear and Tar-Pamlico
watersheds. These two watersheds exhibit a humid subtropical climate, resulting in high
precipitation rates that produce elevated evapotranspiration rates, making the epco parameter
critical in evapotranspiration as it adjusts plant root water uptake under water stress conditions.
The newly added parameter, cn3_swf, in SWAT+ enables the simulation of infiltration and
ponding on land surfaces before runoff commences following a dry period (Wagner et al., 2022).
Studies by Eini et al. (2023), Tumsa et al. (2022), Wagner et al. (2022), and White et al. (2022)
Chapter 1
34
introduced a general view of cn3_swf parameter which accounts for soil water content to adjust
the daily curve number. However, we are lacking complete documentation about the use of
cn3_swf for SWAT+’s calibration. We found that Cape Fear River’s flow is sensitive to cn3_swf
under gridMET and GPM IMERG rainfall datasets which could be explained due to its
characteristics as having humid subtropical climate, resulting in a high amount of rainfall recorded
annually.
In addition, this watershed’s soils are highly porous with permeable sediments that could
increase the total amount of rainfall recharging to groundwater. Besides, the Tar-Pamlico
watershed exhibited high sensitivity to esco (soil evaporation compensation factor) with the GPM
IMERG dataset, contributing to SWAT+’s capability to capture the watershed’s characteristics as
a humid subtropical climatic region with GPM IMERG.
Many SWAT and a few SWAT+ studies mentioned the impacts of watershed
characteristics on the performance of rainfall products (Pulighe et al., 2021; Cho et al., 2009). In
this study, we also found that the sensitivity of calibration parameters is related to the
characteristics of the watershed. Our findings aligned with the studies of Fereidoon et al. (2019)
and Guo & Su. (2019) in Tar-Pamlico and Cape Fear, in which curve number (cn2) was found to
be the most sensitive parameter with all chosen rainfall datasets and watershed characteristics. In
addition, our results aligned with previous works (Migliaccio & Chaubey, 2008; G. Wang et al.,
2014; Zhao et al., 2018) that identified revap_co, flo_min, revap_min, awc, alpha, perco, epco,
cn3_swf, and esco as significant parameters affecting river flow.
4.2. Flow simulation, rainfall dataset and optimization technique
In this study, we analyzed the performance of three rainfall datasets and three optimization
techniques on streamflow prediction. We found GPM IMERG rainfall dataset when used with
Chapter 1
35
GLUE optimization technique generally yields better results for flow simulations. Our findings
partially address the research gaps identified by Yuan et al. (2019), who called for an evaluation
of the compatibility between the GPM IERG dataset and the GLUE optimization technique for
optimal flow simulations. Mousavi et al., 2023 reported high values of KGE (0.88) for IMERG-
Terra model when used with the GLUE optimization technique.
We have found that the selection of a rainfall dataset can play a vital role in flow
simulations. This observation aligns with previous literature on SWAT modeling, where multiple
rainfall datasets were used to assess their impact on optimal flow simulations (Cho et al., 2009;
Dos et al., 2022). We found a couple of different combinations for optimal flow simulations based
on the specific modeling objective. As mentioned earlier, for achieving optimal overall and peak
flow simulation, the combination of GPM IMERG rainfall data with the GLUE optimization
technique has proven to be more effective. For optimal base flow simulation, the combinations of
the ERA-5 rainfall dataset with either DDS or LHS optimization techniques, or the GPM IMERG
rainfall dataset with either GLUE or LHS optimization techniques have been found to work better.
The GPM IMERG rainfall dataset showed the best simulation in terms of peak flow
magnitude compared to other rainfall datasets. Our results align with previous literature, wherein
Tran et al. (2023) indicated that GPM IMERG outperformed five other different satellite
precipitation products for flood peak simulations over eleven Vietnam River Basins using SWAT
model. This product was rated highly in this assessment due to its design utilizing multiple
microwave and infrared sensors, and post-calibration using gauge-based products (Hou et al.,
2014).
In general, SWAT+ performance varied greatly for peak flow magnitude simulation, with
the GPM IMERG rainfall dataset and GLUE optimization technique combination yielding the best
Chapter 1
36
results. SWAT+ underperformed in capturing peak flow duration across various scenarios. Both
SWAT and SWAT+ are designed primarily for long-term continuous predictions and have
limitations when it comes to capturing instantaneous peak flows (Yu et al., 2018). When simulating
flows with all three rainfall products and optimization methods, we consistently experienced an
issue of timing peak flow and the subsequent baseflow recession occurring too early. This
underscores the significant influence of the curve number (cn2) on hydrograph dynamics during
events equivalent to hurricane intensity (Caviedes-Voullième et al., 2012; Lyon et al., 2004). The
accuracy of streamflow simulations displayed notable differences depending on the rainfall
datasets used. This reinforces the observation that in humid climates, SWAT+ flow simulations
exhibit sensitivity to the specific rainfall data (Tan & Yang, 2020).
The Cape Fear watershed is predominantly characterized by sandy soils, conducive to
infiltration, is considered as groundwater-driven watershed (Ewen et al., 2011; Valley, 1988),
leading to pronounced baseflows sourced from a combination of deep, confined, and shallow
aquifers (Ewen et al., 2011; Valley, 1988). However, for the crucial output of baseflow, SWAT+
underperformed in this specific watershed. The model did not accurately predict the magnitude of
baseflow. This could be explained by the model’s simplified groundwater flow equations. These
equations potentially underperform to fully capture the intricacies of baseflow in a watershed that's
so heavily influenced by groundwater, as indicated by Bailey et al. (2020).
Our findings indicate that there’s no single combination of rainfall dataset and optimization
technique is universally superior to other which aligns with previous literature (Dos Santos et al,.
2022; Mararakanye et al., 2020). The combination should be selected based on data availability,
computational power availability, watershed location, and watershed characteristics. GPM
IMERG (Hou et al., 2014) and ERA-5 (Muñoz-sabater et al., 2021) data are available worldwide,
Chapter 1
37
whereas gridMET data is available only for the contiguous United States (Abatzoglou, 2013).
There are not regional limitations for choosing optimization techniques; all of them can be
accessed worldwide through various open-source applications. We compared these techniques
using an equal number of simulations to demonstrate how they perform with similar computational
power.
4.3. Effects of rainfall dataset and optimization technique on performance indices
Performance metrics (NSE, KGE, and R2) are formed differently to measure the goodness-of-fit
of the simulation and could be biased based on data magnitude, outliers, repetition, and other
factors (Ritter & Muñoz-Carpena, 2013). In this study, ERA-5 and GPM IMERG datasets
performed better than the gridMET rainfall dataset when evaluated using KGE and NSE indices
(Figure 1.5). The worst performance index values for gridMET rainfall data was indicated when
using with the DDS technique (NSE, KGE, R2 — Tar-Pamlico: -0.236, 0.057, 0.212 — Cape Fear:
0.029, 0.148, 0.304). Our findings contrast with findings from Tolson and Shoemaker (2007),
which determined that DDS was superior in efficiency compared to GLUE. For DDS, our
experiment was performed with 175 iterations across 12 threads, yielding a total of 2,100
simulations. We hypothesize that the chosen 175 iterations or the 12-thread configuration might
not be adequate to encompass the extensive variety of parameter combinations when employing
the DDS’s second parallel approach, as mentioned by Nguyen et al. (2022). The method, which
involves parallel processing by assigning parameter values to subsequent cores, remains an area
where further validation is needed (Nguyen et al., 2022).
Our findings, as shown in Figure 1.5, revealed that KGE exhibited more consistent
variations across different scenarios, while NSE values showed abrupt fluctuations when assessing
the performance of three rainfall datasets and three optimization techniques in both watersheds.
Chapter 1
38
The robust performance of KGE, especially when compared to NSE, can be attributed to its
complex assessment of hydrological model outputs. Specifically, KGE incorporates three different
components: correlation (r), variability (alpha), and bias (beta) (Knoben et al., 2019). In contrast,
NSE is calculated using the mean squared error between the simulated and observed flows
(Moriasi et al., 2007). KGE showed a better performance in identifying model deficiencies and
improvements (Pool et al., 2018). This more thorough assessment typically translates to
hydrological simulations that are both more precise and dependable (Gupta et al., 2009; Moriasi
et al., 2007). It is worth noting that had we adjusted the model specifically for KGE and R2, similar
to our optimization for NSE, their performance metrics likely would have been higher.
4.4. SWAT+ parameter uncertainty
In this study, we performed experiments with the assumption that the only sources of uncertainty
in predicting these parameters stemmed from the rainfall dataset and the optimization techniques
used. Thus, scenarios are performed with the same set of input (e.g., soil map, land cover, and
observed data) during the modeling process. Our analysis revealed considerable variations,
especially in the flo_min parameters for both Tar-Pamlico (ranging from 0.03 to 0.48) and Cape
Fear (-0.11 to 0.49) as well as in the revap_min parameters for Tar-Pamlico (-0.25 to 0.21) and
Cape Fear (-0.22 to 0.18). The observed variations suggest higher uncertainties in these parameters
compared to the other three that were sensitive in all scenarios, emphasizing the need for cautious
consideration and potential refinement.
The optimization methods for SWAT+ that we examined were successful in
parameterizing cn2 parameter across various HRUs. In SWAT+, groundwater flow parameters,
namely revap_co, and revap_min, are delineated at the aquifer level, allowing for a precise
parameterization (SWAT+, 2020). The minimal variation observed in the awc parameter
Chapter 1
39
highlights SWAT+'s effectiveness in accurately modeling both the flow movement through
landscapes and its absorption by vegetation. It has implications for predictions related to runoff
and evapotranspiration. The minor discrepancy in the cn2 values across the two watersheds might
be attributed to similarities in either land slope or land cover across these regions. Supporting this,
Chordia et al. (2022) identified a lesser degree of variation in cn2 in comparison to eight other
different parameters over four watersheds. In conclusion, SWAT+ parameterizes key components,
such as cn2, the groundwater revap coefficient, and the threshold depth of shallow aquifers where
revap takes place, underscoring its effectiveness in SWAT+ modeling.
This study's limitation lies in its use of a broad calibration approach across two watersheds
without extensively detailing model inputs or fine-tuning parameters for localized hydrological
processes, which could impact the precision of flow predictions in specific scenarios. Additionally,
its focus on comparative analysis restricts the ability to fully optimize the model for any single
dataset or calibration technique.
5. Conclusion
The aim of this study was to comprehensively investigate how the selection of rainfall datasets and
autocalibration techniques affect hydrological model flow simulations. We implemented our
experiments on two major watersheds located in North Carolina and evaluated them using three
performance indices. We found five common sensitive flow parameters — cn2, revap_co, flo_min,
revap_min, awc — under all scenarios, with cn2 being the most sensitive parameter. The newly
added SWAT+ parameter cn3_swf was found to be sensitive in the groundwater-driven watershed.
The sensitivity of the cn3_swf parameter in the groundwater-driven watershed emphasizes the
interconnectedness of surface and subsurface hydrological processes within SWAT+, even when
Chapter 1
40
groundwater dynamics predominate. Significant variation in the simulated flow under different
scenarios of the rainfall dataset and optimization techniques was observed. In most cases, we found
overall better flow predictions for ERA-5 and GMP IMERG rainfall datasets with the GLUE
optimization technique. We found comparatively better results for the gridMET dataset with the
LHS optimization technique.
Autocalibration techniques play an essential role in simulating SWAT+ baseflow
accurately compared to storm flows. The rainfall dataset plays a crucial role in both storm flows
and baseflow predictions. Overall, SWAT+ does better in capturing baseflow magnitude than peak
flow magnitude. GPM IMERG rainfall dataset did a better job than ERA-5 and gridMET in
predicting peak flow magnitude. We suggest the combination of the ERA-5 dataset with DDS or
LHS, the gridMET dataset with LHS, and the GPM IMERG dataset with either GLUE or LHS to
improve baseflow predictions. SWAT+ fell short in predicting baseflow magnitude in the
groundwater-driven watershed. SWAT+ has limited capacity to predict extreme runoff events. We
were not able to adequately simulate the peak runoff duration caused by Hurricane Irene (North
Carolina, Aug 27, 2011) with SWAT+ using any combination of rainfall datasets and optimization
techniques. However, GPM IMERG did a better job in capturing peak runoff magnitude with the
GLUE method than any other combination.
We found some common trends in performance indices among different scenarios, with
NSE giving lower values than KGE and R2 in all cases. Depending on the rainfall dataset and
optimization technique used, either KGE or R2 produced the highest values. We found R2 to
provide consistently higher values than KGE and NSE for the gridMET rainfall dataset. Overall,
we found the GPM IMERG data along with the GLUE optimization technique and KGE
performance index for optimal flow modeling. This study will be helpful for hydrological
Chapter 1
41
modelers to wisely select the rainfall dataset, calibration method, and performance index
combination to predict the more accurate simulated flow. Future studies should explore the effect
of the different number of iterations on flow prediction under different rainfall datasets. A better
technique for estimating minimum aquifer storage to allow return flow is needed in groundwater-
driven watersheds in SWAT+.
Acknowledgments
This research was supported by the Center for Sustainable Energy and Environmental Engineering
and Water Resources Center at East Carolina University. This work was funded by the National
Science Foundation [Funder ID: 2009185 and 2052889] and the Environmental Protection Agency
[Funder ID: R840181]. We are grateful to Drs. Ariane Peralta, Meg Blome, Nancy Sammons, Julie
Miller, and Zeke Holloman for their insightful feedback on this work. We also thank the editors,
Drs. Momcilo Markus and Ying Ouyang, along with the anonymous reviewers, for their
constructive comments.
Data availability
In accordance with the FAIR (Findable, Accessible, Interoperable, and Reusable) principles, we
will make all data, results, and code from our study publicly available in the GitHub repository
https://github.com/EtheridgeLab/Tar_Pam_SWAT following the publication of our findings,
ensuring transparency and accessibility.. The majority of the data utilized in our research is sourced
from open databases.
Chapter 1
42
ORCID
Mahesh R Tapas (https://orcid.org/0000-0001-8833-5531)
References
Abatzoglou, J. T. (2013). Development of gridded surface meteorological data for ecological
applications and modelling. International Journal of Climatology, 33(1), 121–131.
https://doi.org/10.1002/joc.3413
Abatzoglou, J. T., & Ficklin, D. L. (2017). Climatic and physiographic controls of spatial
variability in surface water balance over the contiguous United States using the Budyko
relationship. Water Resources Research, 53(9), 7630–7643.
Adeyeri, O. E., Laux, P., Arnault, J., Lawin, A. E., & Kunstmann, H. (2020). Conceptual
hydrological model calibration using multi-objective optimization techniques over the
transboundary Komadugu-Yobe basin, Lake Chad Area, West Africa. Journal of Hydrology:
Regional Studies, 27, 100655.
Ahmed, Z., Tran, T. N. D., & Nguyen, Q. B. (2020). Applying semi distribution hydrological
model SWAT to assess hydrological regime in Lai Giang catchment, Binh Dinh Province,
Vietnam. Proceedings of the 2nd Conference on Sustainability in Civil Engineering
(CSCE’20), Capital University of Science and Technology, Islamabad, Pakistan, 8.
https://csce.cust.edu.pk/archive/20-404.pdf
Anaba, L. A., Banadda, N., Kiggundu, N., Wanyama, J., Engel, B., & Moriasi, D. (2017).
Application of SWAT to Assess the Effects of Land Use Change in the Murchison Bay
Chapter 1
43
Catchment in Uganda. Computational Water, Energy, and Environmental Engineering, 06(01),
24–40. https://doi.org/10.4236/cweee.2017.61003
Arnold JG, Srinivasan R, Muttiah RS, Williams Jr. 1998. Large area hydrologic modeling and
assessment part I: model development. Journal of the American Water Resources Association
34(1): 73–89
Aryal, A., Tran, T. N. D., Kim, K. Y., Rajaram, H., & Lakshmi, V. (Venkat). (2022). Climate and
Land Use/Land Cover Change Impacts on Hydrological Processes in the Mountain Watershed
of Gandaki River Basin, Nepal. AGU Fall Meeting Abstracts, 2022, H52L-0615.
Bailey, R. T., Park, S., Bieger, K., Arnold, J. G., & Allen, P. M. (2020). Enhancing SWAT+
simulation of groundwater flow and groundwater-surface water interactions using
MODFLOW routines. Environmental Modelling & Software, 126, 104660.
Barrett, J. P. (1974). The coefficient of determination—some limitations. The American
Statistician, 28(1), 19–20.
Betrie, G. D., Mohamed, Y. A., van Griensven, A., & Srinivasan, R. (2011). Sediment
management modelling in the Blue Nile Basin using SWAT model. Hydrology and Earth
System Sciences, 15(3), 807–818.
Beven, K., & Binley, A. (1992). The future of distributed models: model calibration and
uncertainty prediction. Hydrological Processes, 6(3), 279–298.
Beven, K., & Freer, J. (2001). Equifinality, data assimilation, and uncertainty estimation in
mechanistic modelling of complex environmental systems using the GLUE methodology.
Journal of Hydrology, 249(1–4), 11–29.
Chapter 1
44
Bieger, K., Arnold, J. G., Rathjens, H., White, M. J., Bosch, D. D., Allen, P. M., Volk, M., &
Srinivasan, R. (2017). Introduction to SWAT+, a completely restructured version of the soil
and water assessment tool. JAWRA Journal of the American Water Resources Association,
53(1), 115–130.
Blasone, R.-S., Vrugt, J. A., Madsen, H., Rosbjerg, D., Robinson, B. A., & Zyvoloski, G. A.
(2008). Generalized likelihood uncertainty estimation (GLUE) using adaptive Markov Chain
Monte Carlo sampling. Advances in Water Resources, 31(4), 630–648.
Brauman, K. A., Bremer, L. L., Hamel, P., Ochoa‐Tocachi, B. F., Roman‐Dañobeytia, F.,
Bonnesoeur, V., Arapa, E., & Gammie, G. (2022). Producing valuable information from
hydrologic models of nature‐based solutions for water. Integrated Environmental Assessment
and Management, 18(1), 135–147.
C3S. (2017). ERA5: Fifth generation of ECMWF atmospheric reanalyses of the global climate.
European Centre for Medium-Range Weather Forecasts.
https://cds.climate.copernicus.eu/cdsapp#!/home
Caviedes-Voullième, D., García-Navarro, P., & Murillo, J. (2012). Influence of mesh structure on
2D full shallow water equations and SCS Curve Number simulation of rainfall/runoff events.
Journal of Hydrology, 448, 39–59.
Celray, J. (2022). SWAT+ Toolbox - Home.
Chaplot, V. (2014). Impact of spatial input data resolution on hydrological and erosion modeling:
Recommendations from a global assessment. Physics and Chemistry of the Earth, Parts A/B/C,
67, 23–35.
Chapter 1
45
Chordia, J., Panikkar, U. R., Srivastav, R., & Shaik, R. U. (2022). Uncertainties in Prediction of
Streamflows Using SWAT Model—Role of Remote Sensing and Precipitation Sources.
Remote Sensing, 14(21), 5385.
Pulighe, G., Lupia, F., Chen, H., & Yin, H. (2021). Modeling climate change impacts on water
balance of a Mediterranean watershed using SWAT+. Hydrology, 8(4), 157.
Cho, J., Bosch, D., Lowrance, R., Strickland, T., & Vellidis, G. (2009). Effect of spatial
distribution of rainfall on temporal and spatial uncertainty of SWAT output. Transactions of
the ASABE, 52(5), 1545-1556.
Cooper, V. A. (2007). Calibration of conceptual rainfall – runoff models using global optimisation
methods with hydrologic process-based parameter constraints. Journal of Hydrology, 334,
455–466. https://doi.org/10.1016/j.jhydrol.2006.10.036
Chang, C., Lee, H., Do, S. K., Du, T. L. T., Markert, K., Hossain, F., Khalique, S., Piman, T.,
Meechaiya, C., Bui, D. D., Bolten, J. D., Hwang, E., & Chul, H. (2023). Operational
forecasting inundation extents using REOF analysis (FIER) over lower Mekong and its
potential economic impact on agriculture. Environmental Modelling and Software,
162(August 2022), 105643. https://doi.org/10.1016/j.envsoft.2023.105643
Dile, Y., Srinivasan, R., & George, C. (2019). QGIS Interface for SWAT+(QSWAT+), version
1.2. 2. Texas AM University.
Dos Santos, V., Oliveira, R. J., Datok, P., Sauvage, S., Paris, A., Gosset, M., & Sánchez-Pérez, J.
M. (2022). Evaluating the performance of multiple satellite-based precipitation products in the
Congo River Basin using the SWAT model. Journal of Hydrology: Regional Studies, 42,
101168.
Chapter 1
46
Du, T. L. T., Lee, H., Bui, D. D., Graham, L. P., Darby, S. D., Pechlivanidis, I. G., Leyland, J.,
Biswas, N. K., Choi, G., Batelaan, O., Bui, T. T. P., Do, S. K., Tran, T. V., Nguyen, H. T., &
Hwang, E. (2022). Streamflow Prediction in Highly Regulated, Transboundary Watersheds
Using Multi-Basin Modeling and Remote Sensing Imagery. Water Resources Research, 58(3),
1–25. https://doi.org/10.1029/2021WR031191
Efstratiadis, A., & Koutsoyiannis, D. (2010). One decade of multi-objective calibration approaches
in hydrological modelling: a review. Hydrological Sciences Journal–Journal Des Sciences
Hydrologiques, 55(1), 58–78.
Eini, M. R., Massari, C., & Piniewski, M. (2023). Satellite-based soil moisture enhances the
reliability of agro-hydrological modeling in large transboundary river basins. Science of the
Total Environment, 873, 162396.
Ewen, C. R., Whyte, T. R., & Davis Jr, R. P. S. (2011). The Archaeology of North Carolina: Three
Archaeological Symposia. Carolina Archaeological Council Publication.
Fereidoon, M., Koch, M., & Brocca, L. (2019). Predicting rainfall and runoff through satellite soil
moisture data and SWAT modelling for a poorly gauged basin in Iran. Water, 11(3), 594.
Gassman, P. W., Jeong, J., Boulange, J., Narasimhan, B., Kato, T., Somura, H., Watanabe, H.,
Eguchi, S., Cui, Y., & Sakaguchi, A. (2022). Simulation of rice paddy systems in SWAT: A
review of previous applications and proposed SWAT+ rice paddy module. International
Journal of Agricultural and Biological Engineering, 15(1), 1–24.
Gassman, P. W., Jeong, J., Boulange, J., Narasimhan, B., Kato, T., Somura, H., Watanabe, H.,
Eguchi, S., Cui, Y., Sakaguchi, A., Tu, L. H., Jiang, R., Kim, M. K., Arnold, J. G., Ouyang,
W., Waidler, D., White, M. J. M., Steglich, E., Wang, S., … Tar-Pamlico Basin Oversight
Chapter 1
47
Committee. (2007). 済無No Title No Title No Title. Water Resources Research, 43(6), 1–16.
https://doi.org/10.1016/B978-0-12-818597-1.50012-6
Getirana, A. C. V. (2010). Integrating spatial altimetry data into the automatic calibration of
hydrological models. Journal of Hydrology, 387(3–4), 244–255.
Grimaldi, S., Nardi, F., Piscopia, R., Petroselli, A., & Apollonio, C. (2021). Continuous hydrologic
modelling for design simulation in small and ungauged basins: A step forward and some tests
for its practical use. Journal of Hydrology, 595, 125664.
Guo, J., & Su, X. (2019). Parameter sensitivity analysis of SWAT model for streamflow simulation
with multisource precipitation datasets. Hydrology Research, 50(3), 861–877.
Gupta, A., & Govindaraju, R. S. (2019). Propagation of structural uncertainty in watershed
hydrologic models. Journal of Hydrology, 575, 66–81.
Gupta, H. V., Kling, H., Yilmaz, K. K., & Martinez, G. F. (2009). Decomposition of the mean
squared error and NSE performance criteria: Implications for improving hydrological
modelling. Journal of Hydrology, 377(1–2), 80–91.
https://doi.org/10.1016/j.jhydrol.2009.08.003
Ha, L. T., & Bastiaanssen, W. G. M. (2018). Calibration of Spatially Distributed Hydrological
Processes and Model Parameters in SWAT Using Remote Sensing Data and an Auto-
Calibration Procedure: A Case Study in a Vietnamese River Basin. Water (Switzerland), 10(2).
https://doi.org/10.3390/w10020212
Chapter 1
48
Herman, J. D., Quinn, J. D., Steinschneider, S., Giuliani, M., & Fletcher, S. (2020). Climate
adaptation as a control problem: Review and perspectives on dynamic water resources
planning under uncertainty. Water Resources Research, 56(2), e24389.
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Nicolas, J., Peubey, C., Radu, R.,
Balsamo, G., Bonavita, M., Dee, D., Dragani, R., Flemming, J., Forbes, R., Geer, A., Hogan,
R. J., Janisková, H. M., Keeley, S., Laloyaux, P., … Thépaut, J. (2020). The ERA5 global
reanalysis. Quarterly Journal of the Royal Meteorological Society, 146(June), 1999–2049.
https://doi.org/10.1002/qj.3803
Hou, A. Y., Kakar, R. K., Neeck, S., Azarbarzin, A. A., Kummerow, C. D., Kojima, M., Oki, R.,
Nakamura, K., & Iguchi, T. (2014). The global precipitation measurement mission. Bulletin of
the American Meteorological Society, 95(5), 701–722. https://doi.org/10.1175/BAMS-D-13-
00164.1
Jain, S. K., & Sudheer, K. P. (2008). Fitting of hydrologic models: a close look at the Nash–
Sutcliffe index. Journal of Hydrologic Engineering, 13(10), 981–986.
Jha, M., Arnold, J. G., Gassman, P. W., Giorgi, F., & Gu, R. R. (2006). Climate change sensitivity
assessment on Upper Mississippi River Basin streamflows using SWAT. Journal of the
American Water Resources Association, 42(4), 997–1015. https://doi.org/10.1111/j.1752-
1688.2006.tb04510.x
Kansara, P., Sen, I., & Lakshmi, V. V. (2021). Application of the SWAT model for nutrient
modelling of the Narmada River basin. AGU Fall Meeting Abstracts, 2021, H15E-1086.
Chapter 1
49
Knoben, W. J. M., Freer, J. E., & Woods, R. A. (2019). Inherent benchmark or not? Comparing
Nash–Sutcliffe and Kling–Gupta efficiency scores. Hydrology and Earth System Sciences,
23(10), 4323–4331.
Kumar, B., Lakshmi, V., & Patra, K. C. (2017). Evaluating the Uncertainties in the SWAT Model
Outputs due to DEM Grid Size and Resampling Techniques in a Large Himalayan River Basin.
Journal of Hydrologic Engineering, 22(9), 04017039. https://doi.org/10.1061/(asce)he.1943-
5584.0001569
Le, M. H., Lakshmi, V., Bolten, J., & Bui, D. Du. (2020). Adequacy of Satellite-derived
Precipitation Estimate for Hydrological Modeling in Vietnam Basins. Journal of Hydrology,
586(March), 124820. https://doi.org/10.1016/j.jhydrol.2020.124820
Lee, J., Kim, J., Jang, W. S., Lim, K. J., & Engel, B. A. (2018). Assessment of baseflow estimates
considering recession characteristics in SWAT. Water, 10(4), 371.
Levy, M. C., Cohn, A., Lopes, A. V., & Thompson, S. E. (2017). Addressing rainfall data selection
uncertainty using connections between rainfall and streamflow. Scientific Reports, 7(1), 1–12.
Li, X., Chen, Y., Deng, X., Zhang, Y., & Chen, L. (2021). Evaluation and hydrological utility of
the GPM IMERG precipitation products over the Xinfengjiang River Reservoir basin, China.
Remote Sensing, 13(5), 866.
Lovett, E. N. N. R. N. J. C. (2018). Effect of single and multi-site calibration techniques on
hydrological model performance , parameter estimation and predictive uncertainty: a case
study in the Logone catchment, Lake Chad basin. Stochastic Environmental Research and Risk
Assessment, 32(6), 1665–1682. https://doi.org/10.1007/s00477-017-1466-0
Chapter 1
50
Luo, Y., Arnold, J., Allen, P., & Chen, X. (2012). Baseflow simulation using SWAT model in an
inland river basin in Tianshan Mountains, Northwest China. Hydrology and Earth System
Sciences, 16(4), 1259–1267.
Lyon, S. W., Walter, M. T., Gérard‐Marchant, P., & Steenhuis, T. S. (2004). Using a topographic
index to distribute variable source area runoff predicted with the SCS curve‐number equation.
Hydrological Processes, 18(15), 2757–2771.
Maviza, A., & Ahmed, F. (2021). Climate change/variability and hydrological modelling studies
in Zimbabwe: a review of progress and knowledge gaps. SN Applied Sciences, 3(5), 549.
Mararakanye, N., Le Roux, J. J., & Franke, A. C. (2020). Using satellite-based weather data as
input to SWAT in a data poor catchment. Physics and Chemistry of the Earth, Parts A/B/C,
117, 102871.
Migliaccio, K. W., & Chaubey, I. (2008). Spatial distributions and stochastic parameter influences
on SWAT flow and sediment predictions. Journal of Hydrologic Engineering, 13(4), 258–269.
Mirzaei, M., Huang, Y. F., El-Shafie, A., & Shatirah, A. (2015). Application of the generalized
likelihood uncertainty estimation (GLUE) approach for assessing uncertainty in hydrological
models: a review. Stochastic Environmental Research and Risk Assessment, 29, 1265–1273.
Mohammed, I. N., Bolten, J. D., Srinivasan, R., & Lakshmi, V. (2018a). Improved hydrological
decision support system for the Lower Mekong River Basin using satellite-based earth
observations. Remote Sensing, 10(6). https://doi.org/10.3390/rs10060885
Chapter 1
51
Mohammed, I. N., Bolten, J. D., Srinivasan, R., & Lakshmi, V. (2018b). Satellite observations and
modeling to understand the Lower Mekong River Basin streamflow variability. Journal of
Hydrology, 564(January), 559–573. https://doi.org/10.1016/j.jhydrol.2018.07.030
Mondal, A., Le, M. H., & Lakshmi, V. (2022). Land use, climate, and water change in the
Vietnamese Mekong Delta (VMD) using earth observation and hydrological modeling. Journal
of Hydrology: Regional Studies, 42(June), 101132. https://doi.org/10.1016/j.ejrh.2022.101132
Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., & Veith, T. L.
(2007). Model evaluation guidelines for systematic quantification of accuracy in watershed
simulations. Transactions of the ASABE, 50(3), 885–900.
Mousavi, R., Nasseri, M., Abbasi, S., Taheri, M., & Shamsi Anboohi, M. (2023). Global gridded
products efficiency in closing water balance models: various modeling scenarios for behavioral
assessments. Acta Geophysica, 71(5), 2401-2422.
Muñoz-sabater, J., Dutra, E., Agustí-panareda, A., Albergel, C., Hersbach, H., Martens, B.,
Miralles, D. G., Piles, M., & Rodríguez-fernández, N. J. (2021). ERA5-Land: a state-of-the-
art global reanalysis dataset for land applications. Earth System Science Data, January 2020,
4349–4383.
Nash, J. E., & Sutcliffe, J. V. (1970). River flow forecasting through conceptual models part I—
A discussion of principles. Journal of Hydrology, 10(3), 282–290.
NCDEQ. (n.d.-a). Cape Fear River Basin | NC DEQ.
NCDEQ. (n.d.-b). Tar-Pamlico River Basin Documents | NC DEQ.
NCDEQ. (2023). Tar-Pamlico River Basin | NC DEQ GIS Data.
Chapter 1
52
Nguyen, T. V., Dietrich, J., Dang, T. D., Tran, D. A., Van Doan, B., Sarrazin, F. J., Abbaspour,
K., & Srinivasan, R. (2022). An interactive graphical interface tool for parameter calibration,
sensitivity analysis, uncertainty analysis, and visualization for the Soil and Water Assessment
Tool. Environmental Modelling and Software, 156(August).
https://doi.org/10.1016/j.envsoft.2022.105497
Nguyen, B. Q., Tran, T., Łukaszewska, M. G.-, Sinicyn, G., & Lakshmi, V. (2022). Assessment
of Urbanization-Induced Land-Use Change and Its Impact on Temperature, Evaporation, and
Humidity in Central Vietnam. 14(21), 3367. https://doi.org/10.3390/w14213367
Nuong, B. T., Linh, B. K., Trang, V. Q., Huyen, P. T., Thi, B., Thao, P., Ngoc, N. T., & Duong,
B. Du. (2022). Applying Fuzzy Analytical Hierarchy Process to Establish Environmental
Sustainability Indicators for Water Resources Srepok River Basin, Vietnam. VNU Journal of
Science: Earth and Environmental Sciences, 38(4), 63–74. https://doi.org/10.25073/2588-
1094/vnuees.4877
Oeurng, C., Cochrane, T. A., Arias, M. E., Shrestha, B., & Piman, T. (2016). Assessment of
changes in riverine nitrate in the Sesan, Srepok and Sekong tributaries of the Lower Mekong
River Basin. Journal of Hydrology: Regional Studies, 8, 95–111.
https://doi.org/10.1016/j.ejrh.2016.07.004
Oo, H. T., Zin, W. W., & Kyi, C. C. T. (2020). Analysis of streamflow response to changing
climate conditions using SWAT model. Civil Engineering Journal, 6(2), 194–209.
Ozer, D. J. (1985). Correlation and the coefficient of determination. Psychological Bulletin, 97(2),
307.
Chapter 1
53
Pandi, D., Kothandaraman, S., & Kuppusamy, M. (2021). Hydrological models: a review.
International Journal of Hydrology Science and Technology, 12(3), 223–242.
Pool, S., Vis, M., & Seibert, J. (2018). Evaluating model performance: towards a non-parametric
variant of the Kling-Gupta efficiency. Hydrological Sciences Journal, 63(13–14), 1941–1953.
Rawls, W. J., Brakensiek, D. L., & Miller, N. (1983). Green-Ampt infiltration parameters from
soils data. Journal of Hydraulic Engineering, 109(1), 62–70.
Ries, K. G., Newson, J. K., Smith, M. J., Guthrie, J. D., Steeves, P. A., Haluska, T. L., Kolb, K.
R., Thompson, R. F., Santoro, R. D., & Vraga, H. W. (2017). StreamStats. In Version 4: US
Geological Survey Fact Sheet USGS Numbered Series (Vol. 2017).
Ritter, A., & Muñoz-Carpena, R. (2013). Performance evaluation of hydrological models:
Statistical significance for reducing subjectivity in goodness-of-fit assessments. Journal of
Hydrology, 480, 33–45. https://doi.org/10.1016/j.jhydrol.2012.12.004
Schuerz, C. (2022). Getting started with SWATplusR • SWATplusR.
Sehgal, V., Sridhar, V., Juran, L., & Ogejo, J. A. (2018). Integrating climate forecasts with the soil
and water assessment tool (SWAT) for high-Resolution hydrologic simulations and forecasts
in the Southeastern US. Sustainability, 10(9), 3079.
Sharifi, E., Eitzinger, J., & Dorigo, W. (2019). Performance of the state-of-the-art gridded
precipitation products over mountainous terrain: A regional study over Austria. Remote
Sensing, 11(17), 1–20. https://doi.org/10.3390/rs11172018
Smigaj, M., Hackney, C. R., Kieu, P., Dang, V. P., Ngoc, N. T., Du, D., Darby, S. E., & Leyland,
J. (2023). Science of the Total Environment Monitoring riverine traffic from space: The
Chapter 1
54
untapped potential of remote sensing for measuring human footprint on inland waterways.
Science of the Total Environment, 860(September 2022), 160363.
https://doi.org/10.1016/j.scitotenv.2022.160363
Srinivasan, R., Zhang, X., & Arnold, J. (2010). SWAT ungauged: hydrological budget and crop
yield predictions in the Upper Mississippi River Basin. Transactions of the ASABE, 53(5),
1533–1546.
Stein, M. (1987). Large sample properties of simulations using Latin hypercube sampling.
Technometrics, 29(2), 143–151.
Sunmin, K. (2021). Real-time flood forecasting with weather radar and distributed hydrological
model. In Water Engineering Modeling and Mathematic Tools (pp. 369–380). Elsevier.
SWAT+. (2020). I/O documentation for SWAT+. Retrieved from https://swatplus.gitbook.io/io-
docs
Tan, M. L., & Yang, X. (2020). Effect of rainfall station density, distribution and missing values
on SWAT outputs in tropical region. Journal of Hydrology, 584, 124660.
Tapas, M., Etheridge, J. R., Howard, G., Lakshmi, V. V., & Tran, T. N. D. (2022a). Development
of a Socio-Hydrological Model for a Coastal Watershed: Using Stakeholders’ Perceptions.
AGU Fall Meeting Abstracts, 2022, H22O--0996.
Tapas, M., Etheridge, R., Miller Julie. Hinckley Brian, H. Z., & Samantha, F. (2022b).
Hydrological Modeling to Forecast Changes in Eastern North Carolina: Implications for
Agriculture, Climate Change, and Fisheries. North Carolina Water Resources Research
Institute Annual Conference, 1(Lightning talk).
Chapter 1
55
Tran, T. N. D., Nguyen, Q. B., & Zeeshan, A. (2021a). Application of Plaxis for Calculating the
Construction Stability and Soft Embankment in Protecting Ha Thanh. 2nd Conference on
Sustainability in Civil Engineering (CSCE) 2020, 202–210.
https://csce.cust.edu.pk/archive/20-613.pdf
Tran, T. N. D., Zeeshan, A., & Vo, N. D. (2021b). APPLICATION OF HYDRODYNAMIC
MODELLING TO ASSESS THE EFFICIENCY OF HURRICANE PROTECTION
MEASURE AT XOM RO DIKE, PHU YEN PROVINCE, VIETNAM. 2nd Conference on
Sustainability in Civil Engineering (CSCE) 2020, 406. https://csce.cust.edu.pk/archive/20-
406.pdf
Tolson, B. A., & Shoemaker, C. A. (2007). Dynamically dimensioned search algorithm for
computationally efficient watershed model calibration. Water Resources Research, 43(1).
Tran, T. N. D., Nguyen, B. Q., Le, M.-H., Lakshmi, V. (Venkat), Bolten, J. D., & Aryal, A.
(2022a). Robustness of Gridded Precipitation Products in Hydrological Assessment for
Vietnam River basins. AGU Fall Meeting Abstracts, 2022, H22M-07.
Tran, T. N. D., Nguyen, Q. B., Vo, N. D., Marshall, R., & Gourbesville, P. (2022b). Assessment
of Terrain Scenario Impacts on Hydrological Simulation with SWAT Model. Application to
Lai Giang Catchment, Vietnam. In Advances in Hydroinformatics (Issue 2022, pp. 1205–
1222). https://doi.org/10.1007/978-981-19-1600-7_77
Tran, T. N. D., Nguyen, Q. B., Nguyen, T. T., Vo, N. D., Nguyen, C. P., & Gourbesville, P.
(2022c). Operational Methodology for the Assessment of Typhoon Waves Characteristics.
Application to Ninh Thuan Province, Vietnam. In Advances in Hydroinformatics (Issue 2022,
pp. 887–902). https://doi.org/10.1007/978-981-19-1600-7_55
Chapter 1
56
Tran, T. N. D, Nguyen, Q. B., Tam, D., Le, L., Nguyen, T. D., Vo, N. D., & Gourbesville, P.
(2022d). Evaluate the Influence of Groynes System on the Hydraulic Regime in the Ha Thanh
River, Binh Dinh Province, Vietnam. In Advances in Hydroinformatics (pp. 241–254).
https://doi.org/https://doi.org/10.1007/978-981-19-1600-7_15
Tran, T.-N.-D., & Lakshmi, V. (2022e). The land use changes impacts on socio-economic drivers
and simulation of surface and groundwater in the Eastern Shore of Virginia, the United States.
AGU Fall Meeting Abstracts, 2022, H42D-1270.
Tran, T.-N.-D., Le, M.-H., Zhang, R., Nguyen, B. Q., Bolten, J. D., & Lakshmi, V. (2023a).
Robustness of gridded precipitation products for vietnam basins using the comprehensive
assessment framework of rainfall. Atmospheric Research, 293(15), 106923.
https://doi.org/https://doi.org/10.1016/j.atmosres.2023.106923
Tran, T.-N.-D., Le, M.-H., Zhang, R., Nguyen, B. Q., Bolten, J. D., & Lakshmi, V. (2023b).
Corrigendum to “Robustness of gridded precipitation products for Vietnam basins using the
comprehensive assessment framework of rainfall.” Atmospheric Research, 294, 106945.
https://doi.org/10.1016/j.atmosres.2023.106945
Tran, T.-N.-D., Nguyen, Q. B., Zhang, R., Aryal, A., Łukaszewska, M.-G., Sinicyn, G., &
Lakshmi, V. (2023c). Quantification of Gridded Precipitation Products for the Streamflow
Simulation on the Mekong River Basin Using Rainfall Assessment Framework: A Case Study
for the Srepok River Subbasin, Central Highland Vietnam. Remote Sensing, 15(4), 1–27.
https://doi.org/10.3390/rs15041030
Tran, T. N. D., Nguyen, Q. B., Vo, N. D., Le, M. H., Nguyen, Q. D., Lakshmi, V., & Bolten, J.
(2023d). Quantification of global Digital Elevation Model (DEM) – A case study of the newly
Chapter 1
57
released NASADEM for a river basin in Central Vietnam. Journal of Hydrology: Regional
Studies, 45(October 2022), 101282. https://doi.org/10.1016/j.ejrh.2022.101282
Tran, T.-N.-D., Do, S. K., Nguyen, B. Q., Tran, V. N., Grodzka-Lukaszewska, M., Sinicyn, G., &
Lakshmi, V. (2024). Investigating the Future Flood and Drought Shifts in the Transboundary
Srepok River basin Using CMIP6 Projections. IEEE Journal of Selected Topics in Applied
Earth Observations and Remote Sensing, 17, 7516–7529.
https://doi.org/10.1109/JSTARS.2024.3380514
Tumsa, B. C., Kenea, G., & Tola, B. (2022). The Application of SWAT+ model to quantify the
impacts of Sensitive LULC changes on water balance in Guder catchment, Oromia, Ethiopia.
Heliyon, e12569.
Valley, L. C. F. R. (1988). Geology and tectonic history of the lower Cape Fear River valley,
southeastern North Carolina.
Wagner, P. D., Bieger, K., Arnold, J. G., & Fohrer, N. (2022). Representation of hydrological
processes in a rural lowland catchment in Northern Germany using SWAT and SWAT+.
Hydrological Processes, 36(5), e14589.
Wang, G., Barber, M. E., Chen, S., & Wu, J. Q. (2014). SWAT modeling with uncertainty and
cluster analyses of tillage impacts on hydrological processes. Stochastic Environmental
Research and Risk Assessment, 28, 225–238.
Wang, Y., Jiang, R., Xie, J., Zhao, Y., Yan, D., & Yang, S. (2019). Soil and water assessment tool
(SWAT) model: A systemic review. Journal of Coastal Research, 93(SI), 22–30.
Chapter 1
58
Wealands, S. R., Grayson, R. B., & Walker, J. P. (2005). Quantitative comparison of spatial fields
for hydrological model assessment––some promising approaches. Advances in Water
Resources, 28(1), 15–32.
Wei, C., Dong, X., Ma, Y., Gou, J., Li, L., Bo, H., Yu, D., & Su, B. (2023). Applicability
comparison of various precipitation products of long-term hydrological simulations and their
impact on parameter sensitivity. Journal of Hydrology, 618, 129187.
Wellen, C., Kamran-Disfani, A.-R., & Arhonditsis, G. B. (2015). Evaluation of the current state
of distributed watershed nutrient water quality modeling. Environmental Science &
Technology, 49(6), 3278–3290.
White, M. J., Arnold, J. G., Bieger, K., Allen, P. M., Gao, J., Čerkasova, N., Gambone, M., Park,
S., Bosch, D. D., & Yen, H. (2022). Development of a Field Scale SWAT+ Modeling
Framework for the Contiguous US. JAWRA Journal of the American Water Resources
Association.
Wu, T., Zhu, L.-J., Shen, S., Zhu, A.-X., Shi, M., & Qin, C.-Z. (2023). Identification of watershed
priority management areas based on landscape positions: An implementation using SWAT+.
Journal of Hydrology, 619, 129281.
WWEE. (2023). SWATplus-CUP.
Yen, H., Park, S., Arnold, J. G., Srinivasan, R., Chawanda, C. J., Wang, R., Feng, Q., Wu, J., Miao,
C., & Bieger, K. (2019). IPEAT+: A built-in optimization and automatic calibration tool of
SWAT+. Water, 11(8), 1681.
Chapter 1
59
Yu, D., Xie, P., Dong, X., Hu, X., Liu, J., Li, Y., Peng, T., Ma, H., Wang, K., & Xu, S. (2018).
Improvement of the SWAT model for event-based flood simulation on a sub-daily timescale.
Hydrology and Earth System Sciences, 22(9), 5001–5019.
Yuan, F., Zhang, L., Soe, K. M. W., Ren, L., Zhao, C., Zhu, Y., ... & Liu, Y. (2019). Applications
of TRMM-and GPM-era multiple-satellite precipitation products for flood simulations at sub-
daily scales in a sparsely gauged watershed in Myanmar. Remote Sensing, 11(2), 140.
Yue, W., Yao, Y., Su, M., Rong, Q., & Xu, C. (2023). Identifying distributions of urban ecosystem
health based on Latin-hypercube sampling and multi-criteria decision analysis framework.
Ecological Indicators, 147, 109957.
Zhang, A., Li, T., Si, Y., Liu, R., Shi, H., Li, X., Li, J., & Wu, X. (2016). Double-layer
parallelization for hydrological model calibration on HPC systems. Journal of Hydrology, 535,
737–747. https://doi.org/10.1016/j.jhydrol.2016.01.024
Zhao, F., Wu, Y., Qiu, L., Sun, Y., Sun, L., Li, Q., Niu, J., & Wang, G. (2018). Parameter
uncertainty analysis of the SWAT model in a mountain-loess transitional watershed on the
Chinese Loess Plateau. Water, 10(6), 690.
Chapter 2: Simulating the Effects of Sea Level Rise on Nitrogen Export Using the SWAT+
Model: A Case Study of the Tar-Pamlico River Basin, NC, USA
Mahesh R Tapasa*, Randall Etheridgeb, Thanh-Nhan-Duc Tranc, Colin Finlayd, Ariane Peraltad,
Natasha Belle, Yicheng Xua, Venkataraman Lakshmic
a Integrated Coastal Program, East Carolina University, Greenville, NC 27858, USA
b Department of Engineering, Center for Sustainable Energy and Environmental Engineering, East
Carolina University, Greenville, NC 27858, USA
c Department of Civil & Environment Engineering, University of Virginia, Charlottesville, VA
22904, USA
d Department of Biology, East Carolina University, NC 27858, USA
e Department of Biological Systems Engineering, Virginia Tech, Blacksburg, VA 24061, USA
*Corresponding author: tapasm21@students.ecu.edu (Mahesh R Tapas).
ORCID: 0000-0001-8833-5531
Abstract:
This study addresses the urgent need to understand the impacts of climate change on coastal
ecosystems by using the SWAT+ model to assess how sea level rise (SLR) affects nitrate export
Chapter 2
61
in a coastal watershed. Our framework for incorporating SLR in the SWAT+ model includes: (1)
reclassifying current land uses to water for areas with elevations below 0.3 meters based on
projections for mid-century; (2) creating new SLR-influenced land use, SLR-influenced crop
database, and hydrological response units for areas with elevations below 2.4 meters; and (3)
adjusting SWAT+ parameters for the SLR-influenced areas to simulate the effects of saltwater
intrusion on processes such as plant yield and denitrification rates.
We demonstrate this approach in the Tar-Pamlico River basin, a coastal watershed in eastern North
Carolina. We calibrated the model for nitrate load at Washington, achieving a Nash-Sutcliffe
Efficiency (NSE) of 0.61. Our findings show SLR significantly alters nitrate delivery to the estuary
through enhanced nitrate mobility throughout the year. We observed an increase in mean annual
nitrate loads from 155,000 kg N under baseline conditions to 157,000 kg N under SLR scenarios,
a change confirmed by a statistically significant paired t-test (p = 2.16×10−10). This pioneering
framework sets the stage for more sophisticated and accurate modeling of SLR impacts in diverse
hydrological scenarios, offering a vital tool for hydrological modelers.
Study region:
Tar-Pamlico River basin, Eastern North Carolina, United States.
Keywords:
SWAT+; Water quality; Nitrate; Coastal Watershed; Sea Level Rise.
Chapter 2
62
Highlights:
1. The study enhances the SWAT+ model to assess how sea level rise impacts nitrate
dynamics in the Tar-Pamlico River basin, NC, USA.
2. The research alters SWAT+ hydrological parameters and inputs to model the primary and
secondary impacts of sea level rise by integrating new SLR-influenced land use, HRUs,
and crop databases.
3. Results show a significant increase in nitrate loads due to sea level rise, emphasizing the
importance of integrating climate change into water quality planning.
Graphical Abstract:
Chapter 2
63
1. Introduction
Water quality is an important aspect of water resources management (Abbaspour et al., 2007).
Since the 20th century, elevated nitrate levels in water bodies have posed significant threats to
coastal watersheds (Pringle 2001; Vilmin et al., 2018). While the invention of nitrogen-based
fertilizers was tremendously advantageous for mass food production, it poses a significant risk to
our coastal ecosystems (Galloway et al., 2013; Fixen & West, 2002). To be specific, high nitrate
levels in water can create an excessive growth of algae and certain aquatic plants. These algal
blooms can hinder sunlight from reaching the bottom of rivers, streams, lakes, or estuaries; leading
to the inhibition of benthic aquatic plant growth (Wurtsbaugh et al., 2019). Eventually, as these
plants and algae die, they are broken down by microbes that consume oxygen (Jewell & McCarty,
1971; Cui et al., 2021), leading to the creation of low-oxygen zones where few organisms can
survive (Gerloff & Krombholz, 1966; Seibel, 2011). Addressing elevated nitrate levels in coastal
watersheds—a global issue—costs billions annually, particularly straining developing countries
and threatening economic stability and aquatic ecosystem health (Sekhon, 1995; Bernhardt et al.,
2005; Craswell 2021; Rasiah et al., 2005; Mathewson et al., 2020). This raises global concern and
should be a major part of a country’s plan for conserving or restoring healthy aquatic ecosystems.
In this study, ecosystem health is defined as the ability of a coastal watershed to sustain biological
productivity (Sherman 1994), maintain ecological processes (O’Brien et al., 2016), support
biodiversity (Qian et al., 2023), and meet societal needs (Christensen et al., 1996). Nitrate loading
to the estuary was used as an indicator of ecosystem health.
Nitrogen gas (N2) makes up the majority of Earth’s atmosphere, representing 78 percent of
its total composition (Hart 1978). Although nitrogen is abundant in the atmosphere, it needs to be
converted to ammonia/ammonium (NH3+/NH4+) through fixation before plants and animals can
Chapter 2
64
use it (Postgate 1998). Bacteria that fix nitrogen are essential as they transform atmospheric
nitrogen gas into forms like ammonium (NH4+) or ammonia (NH3) (Postgate 1998). Animals
obtain nitrogen through eating plants (Temperton et al., 2007), integrating it into the broader food
web (Meunier et al., 2016). When plants and animals die, the organic nitrogen they contain is
broken down by microbes, turning it back into NH4+ (mineralization) and eventually nitrate (NO3-
) via nitrification (Abatenh et al., 2018; Gupta et al., 2017). In addition, denitrifying microbes
assist in converting nitrate back into nitrogen gas or nitrous oxide (N2O) (Vilar-Sanz et al., 2013).
Climate change and anthropogenic activities greatly affect the nitrogen cycle (Aryal et al.,
2022; Bennett et al., 2014; Vitousek et al., 1997), altering the rates of nitrogen fixation (Galloway,
1998), mineralization, and denitrification (Zhu et al., 2015). The alterations in the nitrogen cycle
and nutrient availability impact coastal watersheds, especially where nitrogen is the primary
limiting nutrient, leading to challenges for the health of aquatic ecosystems (Rabalais, 2002). The
nitrogen cycle, carbon cycle, and climate are anticipated to be further altered due to human effects
on the planet (Bernal et al., 2012). Sea levels are rising due to multiple influences of rising
temperatures. (Karl et al., 2009; Cazenave & Cozannet, 2014). Coastal communities are more
susceptible to the effects of climate change (Oliver-Smith, 2009). Specifically, SLR can cause
flooding in low-lying areas, saltwater intrusion, and pose significant threats to aquatic ecosystems
(Knighton et al., 1991). Sea level rise can alter nitrogen dynamics in watersheds by influencing
both the release of ammonium from sediments and the microbial processes of nitrification and
denitrification, impacting nitrogen uptake by plants and overall nutrient retention. Increased
salinity due to SLR can enhance the bioavailability of ammonium, altering plant-microbe
interactions and the nitrogen cycle (Ardón et al., 2013; Kirwan & Megonigal, 2013; Waldron et
al., 1997). For farmers in coastal regions, saltwater intrusion is a pronounced concern due to its
Chapter 2
65
threat to crop growth. The boundary between saltwater and freshwater shifts as the inflow and
outflow rates vary, with an increase in saltwater causing this interface to move further upstream
(Michael et al., 2005). This progression leads to saltwater intrusion, resulting in elevated soil
salinity, which disturbs the equilibrium of the agroecosystem, including crop growth and nitrogen
cycle. This increased salinity, in turn, inhibits the activities of nitrogen-fixing bacteria and
interferes with the conversion of organic nitrogen compounds into forms that plants can easily
assimilate (Etesami & Adl, 2021; Kirova & Kocheva, 2021). Consequently, this disruption in the
nitrogen cycle ultimately reduces nutrient availability for crops (specifically for soybean, which is
a major crop type in the Tar-Pamlico basin), thereby affecting their overall health and yield.
Additionally, high soil salinity can reduce growth or kill crops despite abundant nitrogen
availability (Van et al., 1999), leading to an accumulation of unused nitrogen in soils where
fertilizers are applied. SLR compromises the fertility of the soil in coastal agricultural watersheds
by disrupting nutrient dynamics (Weissman & Tully, 2020). The effects of SLR are not limited to
agriculture, as a higher SLR rate could lead to the drowning of wetlands, reducing their nitrogen
removal potential (Kirwan & Megonigal, 2013). It also would increase the flood frequency, in
which urban areas tend to have higher runoff, resulting from the increase of anthropogenic
nutrients (Macías-Tapia et al., 2021).
Hydrological models are frequently used to assess the potential impacts of climate change
and other anthropogenic influences on nitrogen transport and retention in watersheds (Hattermann
et al., 2006). The Soil and Water Assessment Tool (SWAT) is one of the most widely used models
for this purpose. Developed in the early 1990s, it is a semi-distributed hydrological model
extensively used for water quality (Abbaspour et al., 2007) and quantity analyses (Zhang et al.,
2011). SWAT has proven effective in modeling nitrate loads in diverse watersheds, with
Chapter 2
66
applications in places such as the Vamanapuram River Basin, India (Saravanan et al., 2023),
Brittany, France (Conan et al., 2023), the Lower Seyhan Plain, Turkey (Donmez et al., 2020), and
the Des Moines River, United States (Schilling and Wolter 2009). In recent years, the release of
SWAT+, an enhanced version of SWAT, has gathered attention for its advancements in
understanding interactions within watersheds and sub-watersheds, as well as its improved data
management, analysis, and visualization capabilities (SWAT+, 2020). However, there is a gap in
examining the integration of SLR with the SWAT+ model, possibly due to the model’s limited
ability to manage the bidirectional flow (Bieger et al., 2017).
In this study, our primary objective is to develop a method for simulating the effects of
SLR on nitrogen processing within coastal watersheds using the SWAT+ model. We achieved this
by identifying key parameters and input changes in the model, which enabled simulation of
changes in nitrate transport and retention based on current literature. Additionally, we investigated
how SLR influences nitrate loads in the downstream boundaries of a watershed. The Tar-Pamlico
River Basin, located in eastern North Carolina, is used as a case study, given its historical
challenges with both SLR and elevated nitrate levels over recent decades (Helmers et al., 2022;
Ury et al., 2021). Through this work, we aim to improve the traditional coastal hydrological
modeling framework by partially integrating the complex interactions and changes that arise from
shifting nitrogen cycle due to SLR.
1.1 Objectives
The objectives of this study are to: 1) Develop and optimize a nitrate model for the Tar-Pamlico
watershed using the SWAT+ hydrological model. 2) Conduct a literature search on the effects of
Chapter 2
67
salinity on watershed processes, with a focus on agricultural land use. 3) Develop a novel approach
to incorporating SLR-influenced land use, plant database, and HRUs in SWAT+. 4) Investigate
the effects of SLR on ecosystem health, specifically focusing on changes in nitrate load to the
Pamlico Estuary under baseline and SLR scenarios.
2. Materials and methods
2.1 Study area
This study focuses on the Tar-Pamlico River basin, a coastal watershed in North Carolina,
USA (Figure 2.1). As the fourth-largest watershed in the state, the Tar-Pamlico River basin stands
out as one of just four entirely contained within North Carolina, alongside the Cape Fear, Neuse,
and White Oak River basins (NCDEQ, 2023). With its waters ultimately flowing into the Pamlico
Sound, this watershed has a rich diversity of ecosystems and varied habitats (North Carolina
Department of Environmental Quality (NCDEQ, 2023). The Tar-Pamlico covers an expansive
6,400 square miles and spans 15 counties. It is home to a population exceeding 470,000 residents.
The Tar-Pamlico watershed is divided into five major sub-watersheds, including the Upper Tar
River, Fishing Creek, Lower Tar River, Pamlico River, and Pamlico Sound. The land use within
this watershed is divided among agriculture (27.9%), forests (33.9%), wetlands (31.9%),
pastureland (3.5%), rangeland (1.3%), and urban areas (1.4%) (Claggett et al., 2015). The
freshwater streams and rivers within the basin originate in the Piedmont region in north-central
North Carolina. These waterways flow southeastward and, upon nearing tidal zones, transform
into the expansive (Figure 2.1), tidally influenced estuary (Keith, 2014), enhancing its ecological
complexity and economic productivity (NC DEQ, 2009, 1994).
Chapter 2
68
Figure 2.1 (a) Location of the Tar-Pamlico River basin within the United States, (b) Digital
elevation model and geographical characteristics of the Tar-Pamlico watershed, (c) Hydrological
monitoring station in Washington used for model calibration, (d) Tar-Pamlico’s coastal region
where we incorporated the effects of SLR
2.2. SWAT+ model setup
In this study, we implemented the SWAT+ model (SWAT+ IO Document, 2020) to simulate
hydrological processes throughout the river basin. This tool excels at assessing dynamics within
both watersheds and sub-watersheds. SWAT+ enables the evaluation of how various hydrological
parameters influence watershed dynamics, affecting water quality, agricultural yields, and nitrogen
Chapter 2
69
cycle processes (SWAT+ IO Document, 2020). It allows for detailed simulation of environmental
interactions, aiding the creation of policies and strategies especially aimed at mitigating issues
related to water quality, water availability, and flood risks (SWAT+ IO Document, 2020).
For setting up the SWAT+ model, a variety of data sources were used to accurately depict
watershed characteristics. We obtained watershed boundary data from the United States
Geological Survey (USGS) StreamStats service and incorporated elevation, land use, and soil type
data from the USGS for 2011, the National Land Cover Database (NLCD) for 2008, and the Soil
Survey Geographic Database (SSURGO) for 2015, all at a 90m resolution. These inputs were
processed using QSWAT+ to delineate drainage networks, sub-basins, and Hydrological Response
Units (HRUs), crucial for modeling the intricate hydrological behaviors within the watershed. We
used weather data from the Global Precipitation Measurement Integrated Multi-satellite
Retrievals. Additionally, we assimilated agricultural data, including crop types and fertilizer
application rates, from reports by the North Carolina Department of Environmental Quality
(NCDEQ) and the North Carolina Department of Agriculture & Consumer Services (NCDEQ,
2014; NCGAR, 2022), specifically pertaining to the Tar-Pamlico watershed. This information
proved essential for more accurate simulations of agricultural runoff and nutrient transport.
We also integrated wastewater treatment data for the 22 discharges (Appendix 1) within the
Tar-Pamlico watershed from the National Pollutant Discharge Elimination System (NPDES) and
atmospheric deposition data from the National Atmospheric Deposition Program (NADP, 2023).
Observed flow data were collected from USGS monitoring stations in Greenville and Washington.
Due to incomplete data at the Washington station post-September 2006, we extrapolated the flow
data from Greenville using a linear regression model. This model showed a strong correlation (R²
= 0.90), indicating that approximately 90% of the discharge variability at Washington could be
Chapter 2
70
reliably predicted from Greenville’s data. We used this relationship to compensate for missing
flow data at Washington. Nitrate concentration data was obtained from NCDEQ (site
21NC01WQ).
To calculate monthly nitrate load, we first calculated the average daily discharge at
Washington for each month. Then, we calculated the average nitrate concentration for each month
from the available nitrate concentration data. We multiplied the average discharge for a specific
month by the average nitrate concentration for the same month and then multiplied it by the total
number of days in that month (Preston et al., 1989). In 2003, there were 357 daily observations,
indicating a high-frequency data collection effort at this location (site 21NC01WQ) in Washington,
NC. However, this frequency declined sharply over the years. By 2005, the frequency of data
collection had decreased, with only 144 daily observations. This trend continued, and by 2010,
only 40 daily observations were recorded. Eventually, the data collection frequency shifted to a
monthly scale, with only about 10 to 13 observations per year from 2011 onwards. This reduction
in data collection frequency over time introduces additional challenges in accurately estimating
nitrate loads and concentrations, as fewer data points are available to capture the variability and
trends (Birgand et al., 2011).
2.3. SWAT+ model optimization
We calibrated the model for nitrate load with observed data from Washington, NC (Figure 2.1).
We utilized the SWATrunR package (Schuerz 2019) to calibrate the model. Our calibration efforts
targeted 23 parameters identified as crucial for accurately representing hydrological and nitrogen
processes, based on sensitivity analysis and literature reviews outlined in Appendix 1. For
parameter adjustment, we employed three methods: absolute change (x' = x + y), percent change
(x' = y * x / 100), and absolute value (x' = y), where x is the default value, x' the new value, and y
Chapter 2
71
the calibrated parameter value (SWAT+ IO Document, 2020). The choice of method was
influenced by each parameter's resolution and initial range; for instance, basin-level parameters
predominantly used absolute value adjustments. The decision between absolute and percentage
changes depended on the magnitude of the parameter’s range—absolute changes were preferred
for narrow ranges, while percentage changes were favored for wider ranges. This approach allowed
for a broad range of parameter adjustments during model calibration, ensuring flexibility across
varying resolutions and initial ranges. The detailed optimization table, which includes the
parameter name, type of change, parameter description, and its default and calibrated values, is
mentioned in Appendix 1. We also soft-calibrated the model for yield and denitrification rate for
selected HRUs representing corn, cotton, soybean, and general agricultural crops (Etheridge et al.,
2014). In this study, the general agricultural crop type was used for all other crops (e.g. tobacco,
sweet potatoes, etc.), and in SWAT+, we simulated that using agricultural_land_row (agrr) crop
type. The finalized values of parameter calibration and change type are discussed in Appendix 1.
2.4. SLR incorporation:
We aimed to design a framework to simulate the partial effects of SLR in the SWAT+ model with
a focus on the nitrogen cycle. We achieved this by systematically adjusting inputs and parameters
to reflect the dynamic impacts of SLR. While acknowledging that our base model, SWAT+,
primarily functions as a unidirectional rainfall-runoff model and does not typically account for
backflow (SWAT+ IO Document, 2020), we focus on altering expected land uses, enhancing the
model's capability to simulate critical land processes, such as altered denitrification rates and the
reduction of crop yields under SLR conditions.
We utilized mid-century SLR projections from the National Oceanic and Atmospheric
Administration (NOAA, 2022), which forecasts an increase of approximately 0.3 m by 2050.
Chapter 2
72
Based on NOAA's projections, we reclassified all landcover with elevations less than 0.3 m to
water landcover. To account for the effects of saltwater intrusion, we consider any areas with
elevation less than 2.4 m to fall within the risk zone of saltwater intrusion (NOAA, 2022). For
areas with elevations under 2.4 m, we used GIS to process the land use changes by developing a
new category labeled as saltwater-intruded land use for the respective land uses.
After incorporating these new land use types, we added the saltwater-influenced crops to the
SWAT+ plant database (Appendix 1). Subsequently, using the updated land cover map and the
revised SWAT+ database, we generated new Hydrological Response Units (HRUs) for areas
extending up to 2.4 m above the current mean sea level. For the baseline simulation, we retained
the same parameters for the saltwater-intruded crops as their corresponding conventional crops.
We confirmed that SWAT+ produced consistent outputs for these new saltwater-intruded crops as
for their parent crops when no parameters were altered. The creation of these new saltwater-
intruded HRUs allowed us to modify SWAT+ parameters to simulate the impacts of SLR on the
nitrogen cycle.
We changed multiple SWAT+ parameters in our initial efforts to simulate the effects of SLR.
We conducted a literature review to identify processes and their corresponding SWAT+
parameters potentially affected by saltwater intrusion, such as soil pH (Al-Busaidi & Cookson,
2003) and electrical conductivity (Rhoades & Corwin, 1990). Based on our findings, we adjusted
electrical conductivity (ec.sol) and pH (ph.sol) parameters in the model to account for SLR.
However, despite significant alterations to these parameters, we observed no impact on the
simulated denitrification rate and crop yields. This suggests that these parameters are not directly
linked to the outputs we expected them to alter and additional factors within the SWAT+ model
would need to be considered to reflect the effects of SLR.
Chapter 2
73
To address this challenge, we opted for an indirect approach to incorporating SLR effects. We
introduced new soil nutrient layers in our model to better simulate denitrification rates and plant
yield (SWAT+ IO Document, 2020) under SLR. The parameters that were altered were the
coefficient for adjusting concentrations based on depth (exp_co) and the fraction of active soil
humus (fr_hum_act). In preliminary analysis we found increasing the exp_co parameter decreased
denitrification and decreasing fr_hum_act decreased denitrification. Increasing the exp_co
parameter likely reduces denitrification by limiting the depth at which there are high
concentrations of nitrate, while decreasing fr_hum_act reduces denitrification by lowering the
proportion of active humus that supports microbial activity essential for the denitrification process.
We developed two new soil nutrient layers: one for saltwater-intruded agricultural crops and
another for the rest of the saltwater-intruded land use. The soil nutrient layers were applied at the
HRU resolution for the respective land use (SWAT+ IO Document, 2020). The expected decrease
in denitrification was simulated along with an excessive decrease in yield, thus we adopted a
multiparameter approach to further fine-tune denitrification and plant yield by altering other HRU-
specific parameters such as curve number (cn2), plant uptake compensation factor (epco), soil
evaporation compensation factor (esco), available water capacity (awc), plant ET curve number
coefficient (latq_co), and potential maximum leaf area index (lai_pot). We used SWATrunR
(Schuerz et al., 2022) to alter these parameters specifically for saltwater affected HRUs (up to an
elevation of 2.4 m) to adjust plant yield and denitrification rate per literature under SLR scenarios,
which is further discussed in section 3.2. The values of these parameters under freshwater
conditions and saltwater-influenced conditions are discussed in section 3.2. This innovative
framework is intended to serve as a pioneering step for hydrological modelers, setting the stage
for more sophisticated and accurate modeling of SLR impacts.
Chapter 2
74
3. Results and Discussion
3.1 Model optimization
In this study, we simulated monthly nitrate load at the Washington, NC station from 2001 to 2019
with 2 years of warm up period. We calibrated the model from 2003 to 2011 and validated it from
2012 to 2019 (Figure 2.2). During the calibration period, the model showed a good level of
accuracy within the coastal Tar-Pamlico watershed, achieving Nash-Sutcliffe Efficiency (NSE) of
0.61, Coefficient of Determination (R²) of 0.61, and Kling-Gupta Efficiency (KGE) of 0.77. These
metrics suggest a strong agreement between observed and simulated data, indicating that the model
is well-calibrated and capable of simulating the dynamics governing nitrate transport and retention.
We found lower performance indices values for the validation period; the model's performance
metrics declined to a NSE of 0.33, a R² of 0.33, and a KGE of 0.39, indicating moderate accuracy.
This decline likely resulted from uncertainties associated with the change in nitrate concentration
sampling frequency post-2010 (Birgand et al., 2011). Despite the lower metrics, the model still
demonstrated its capacity to provide reasonable nitrate load predictions under varying conditions
(Upadhyay et al., 2022).
Chapter 2
75
Figure 2.2: Monthly nitrate load optimization results
3.2. Modeling the effects of SLR on nitrogen processes
As mentioned in Section 2.5, simulating the ramifications of SLR on watershed processes, with a
particular focus on the adjustment of plant yield and denitrification were not as straightforward as
expected. Saltwater intrusion disrupts normal plant physiological processes, impeding nutrient
uptake and causing osmotic stress in plants, which is reflected in the reduced yields observed
(Okon, 2019; Safdar et al., 2019). Sea level rise can decrease denitrification rates by increasing
waterlogged conditions, reducing oxygen availability, and inhibiting microbial activity (Hamonts
et al., 2013; Mazhar et al., 2022). SLR reduces potential leaf area (by around 30%) by flooding
coastal areas, increasing soil salinity and altering vegetation, leading to sparse foliage (Bond-
Lamberty et al., 2023), which we adjusted using the potential maximum leaf area index (lai_pot)
parameter in SWAT+. We used previous literature to inform changes made to model parameters
(Table 2.1). Even though SWAT+ may not be capable of simulating backflow, the hydrological
Chapter 2
76
parameter changes discussed in this section should enable the model to more accurately simulate
processes in response to SLR when viewed at the monthly or annual scale. It is highly unlikely to
be accurate at the daily time scale. This SLR-calibrated SWAT+ model offers valuable insights
into the intricate interplay between SLR and watershed processes governing nitrate export.
Chapter 2
77
Table 2.1: Parameters altered simulate plant yield and denitrification changes under SLR
Parameter
Potential Parameter
Range
(Unit)
Resolution
Type of
Change
Description
Default
value
Freshwater
landuse
calibration
SLR calibration
range
SLR-
incorporated
landuse
awc
0.01, 1.0
(mm H2O mm-1)
sol
Absolute
Change
The difference in soil water
content between field capacity and
permanent wilting point
0.14
0.081
-0.3, 0.08
-0.021
esco
0.0, 1.0
(-)
hru
Absolute
Change
Soil evaporation compensation
factor which allows modification
of depth distribution to meet soil
evaporative demand, considering
capillary action, crusting, and
cracks.
0.95
-0.069
-0.3, 0.3
0.038
epco
0.0, 1.0
(-)
hru
Absolute
Change
Plant uptake compensation factor
which allows adjustment of water
uptake depth distribution in
0.5
-0.114
-0.3, 0.3
-0.042
Chapter 2
78
response to plant transpiration
demand and soil water availability.
latq_co
0.0, 1.0
(-)
hru
Absolute
Change
Coefficient for the Plant ET curve
number
0.01
0.101
-0.3, 0.3
0.279
cn2
35.0, 95.0
(-)
hru
Percent
Change
Curve number for Condition II
runoff potential.
12.283
13, 20
17.834
Lai_pot
0.5, 10
(m2 m-2)
Plt
Absolute
Value
Potential maximum leaf area index
Corn
Cotton
Soybean
General agricultural crop
6
4
5
3
5
2.5
2.027
3
3.5
1.75
1.419
2.1
Chapter 2
79
3.2.1. Simulated effects on crop yields:
In this study, we simulated the impacts of SLR on the yields of four distinct crop types—general
agricultural, corn, cotton, and soybean—quantified in kilograms per hectare (kg/ha). Figure 2.3
represents an average of a particular crop’s average annual yield for SLR-influenced HRUs.
Following the adjustments made to the parameters in Table 2.1, our model simulated a reduction
in yields for all crops in the ranges found in the literature (Gibson et al., 2021). In our results,
general agriculture yields decreased the most by 36%, from 4,700 to 3,000 kg/ha, followed by
cotton at 36%, from 2,900 to 1,900 kg/ha, soybean at 33%, from 2,700 to 1,800 kg/ha, and corn at
23%, from 8,900 to 6,800 kg/ha. As expected due to calibration, these findings are consistent with
previous literature on the impacts of salinization on crop yields (Gibson et al., 2021).
Figure 2.3: Simulated effects of SLR on annual average crop yields
Chapter 2
80
3.2.2. Simulated effects on denitrification rate
Changes made to the SLR-influenced land uses were also aimed at simulating the effects of
increased salinity and soil water levels associated with SLR on denitrification. The results of these
changes show a significant decline in denitrification across all crop types (Figure 2.4), with
reductions ranging from 59% in corn to 72% in soybeans, matching with previous literature
(Hofstra & Bouwman, 2005; Qian et al., 1997). The decrease in soil nitrate processing capabilities
under SLR conditions, attributed to increased soil salinity inhibiting microbial activity essential
for nitrogen cycling, could further complicate microbial community structures and potentially
reduce denitrification efficiency (Mazhar et al., 2022; Spivak et al., 2019).
Figure 2.4: Impacts of SLR on annual average denitrification [AGRR: General agricultural crop;
COTS: Cotton; SOYB: Soybean]
Chapter 2
81
3.3. Nitrate load to the estuary
We conducted a comprehensive comparison of the monthly nitrate load at the Pamlico Estuary,
contrasting conditions under both baseline and SLR scenarios. We compared the model results for
both SLR and baseline scenarios from January 2003 to December 2019 (Figure 2.5). This analysis
aimed to quantify the variations in nitrate loads across different months, providing a clearer
understanding of how rising sea levels could affect nutrient export in this coastal environment.
Both datasets exhibit significant variability and peaks over time, but the SLR scenario generally
shows a trend of increased nitrate loads compared to the baseline, especially noticeable in the later
years depicted (Figure 2.5).
Figure 2.5 Simulated Nitrate Load For SLR and Baseline Scenarios
The results reveal notable variations in the seasonal nitrate loads when comparing the
baseline with the SLR scenarios (Table 2.2). Winter and spring exhibited higher nitrate loads
Chapter 2
82
compared to summer and fall. This pattern is likely influenced by increased precipitation during
the colder months (Sayemuzzaman & Jha, 2014), which enhances runoff (Moraglia et al., 2022)
and nitrate leaching (Oh & Sankarasubramanian, 2012), whereas the growing vegetation in warmer
months can assimilate more nutrients (Tian et al., 2014), thereby reducing nitrate levels. The data
also showed that nitrate loads consistently increase across all seasons under SLR conditions, with
the most substantial increases during winter and spring (Table 2.2). This observation suggests that
SLR modifies nitrate availability and transport, particularly during peak precipitation periods
(Stuart et al., 2011). The elevation of the sea level can alter the water table (Hay et al., 1990) and
increase the inundation of low-lying coastal areas (Mohd et al.,2018), resulting in greater surface
runoff (Wang et al., 2011) and potentially higher flows during heavy rains (Rotzoll & Fletcher,
2013). These changes facilitate the more rapid movement of nitrate from terrestrial sources into
estuarine and marine environments, underscoring the impact of SLR on nitrate loss (Munksgaard
et al., 2019; Voss et al., 2015; Wang et al., 2017).
Table 2.2 Average seasonal nitrate load from Tar-Pamlico River Basin
Season
Nitrate Load Baseline
(kg NO3-N)
Nitrate Load SLR
(kg NO3-N)
Percentage Change
(%)
Winter
266,000
267,900
0.73%
Spring
183,500
186,000
1.37%
Summer
93,000
94,500
1.60%
Fall
78,400
79,800
1.80%
The percentage change in nitrate loads from the Tar-Pamlico basin into the Pamlico estuary
indicates a higher relative increase in nitrate loads during the fall (1.80%) and summer (1.60%)
Chapter 2
83
compared to spring (1.37%) and winter (0.73%), under SLR scenarios (Table 2.2). This heightened
increase in nitrate loads during summer and fall under SLR scenarios can be attributed to several
interrelated factors (Herbert et al., 2015). Saltwater intrusion due to SLR increases soil salinity
(Bayabil et al., 2021), disrupting nitrogen cycling (Nelson & Zavaleta, 2012), and diminishing the
nitrate uptake by crops in these growing seasons (He et al., 2018), while simultaneously altering
hydrological dynamics that increase nitrate mobility (Donner & Kucharik, 2003). The disruption
of microbial denitrification processes by increased salinity may result in more residual nitrates that
leach into waterways (Arce et al., 2013). Additionally, seasonal agricultural practices, potentially
misaligned with changing environmental conditions, may further contribute to nitrate runoff
(Conrad & Marinos, 2024).
Statistical comparison of the nitrate load from the Tar-Pamlico basin under baseline and SLR
conditions was conducted using a paired t-test (Table 2.3). The test compared two sets of model
output: the mean annual nitrate loads under the baseline condition (mean = 155,200 kg N) and
under the SLR condition (mean = 157,100 kg N), with each set comprising 204 observations (Table
2.3). The Pearson Correlation coefficient of 0.9995 indicates a very high degree of linear
correlation between the paired samples, which is expected since the measurements under both
scenarios pertain to the same time points (Cohen et al., 2009).
Chapter 2
84
Table 2.3 Statistical summary of paired t-test analysis for monthly nitrate loads under baseline
and SLR conditions
Statistical Measure
Value
Mean Nitrate Load (Baseline)
155,200 kg NO3-N
Mean Nitrate Load (SLR)
157,100 kg NO3-N
Pearson Correlation Coefficient
0.9995
Degrees of Freedom (df)
203
t-Statistic
-6.6869
p-Value (Two-tailed)
2.16×10−10
The paired t-test, yielded a t-statistic of -6.6869, reflecting a significant difference between
the means. The negative t-statistic suggests that the nitrate load under SLR conditions is higher
than under baseline conditions. The two-tailed p-value of approximately 2.16×10−10 indicates that
this difference is highly statistically significant, far below the conventional alpha level of 0.05.
The outcome of this t-test provides strong evidence to reject the null hypothesis, which suggested
no mean difference between the nitrate loads under the two conditions. Therefore, we can infer
with high confidence that SLR has a statistically significant effect on simulated nitrate loads in the
Pamlico Estuary. Overall, the interaction between abiotic factors like precipitation and SLR and
biotic factors such as microbial activity underscores the complex nature of nitrate dynamics in
coastal watersheds (Bouwman et al., 2013). Understanding these interactions is crucial for
developing effective management strategies to mitigate the impacts of nutrient loading and protect
aquatic ecosystems in the context of ongoing climate change (Hamilton et al., 2016).
Chapter 2
85
4. Conclusion
This study presents a novel methodology to simulate the impacts of SLR on nitrate cycling and its
effect on ecosystem health, specifically on nitrate export in the Tar-Pamlico River basin, using the
SWAT+ model. We attempted to innovatively integrate adjustments for the effects of SLR on
hydrological parameters and inputs, significantly enhancing the model's ability to project changes
in nitrate loads due to these environmental changes.
To simulate the effects of SLR in the SWAT+ model, we adjusted parameters related to land
use, crop databases, and HRUs to incorporate both primary impacts like land inundation and
secondary impacts such as saltwater intrusion. Specifically, land uses at elevations below 0.3
meters were reclassified to water, and SWAT+ parameters were adjusted for areas below 2.4
meters to model the effects of increased soil salinity on crop yield and denitrification rates. The
findings reveal that SLR significantly influences nitrate loads, with an observed increase in nitrate
transport to the Pamlico Estuary, particularly during wetter seasons. The introduction of SLR into
the SWAT+ model facilitated a detailed analysis of how altered hydrological dynamics and
increased salinity affect agricultural productivity and ecosystem health.
This study acknowledges the limitations of SWAT+, particularly its inability to simulate
backflow, and recognizes that our approach to modeling land processes under SLR conditions may
not be definitive. However, this research represents a crucial first step toward integrating SLR
impacts into the SWAT+ framework, providing a foundation for further refinement and
development by hydrological modelers.
Appendix 1: https://github.com/EtheridgeLab/Tar_Pam_SWAT/blob/main/Appendix1.docx
Chapter 2
86
References
Abatenh, E., Gizaw, B., Tsegaye, Z., & Tefera, G. (2018). Microbial function on climate change-
a review. Environment Pollution and Climate Change, 2(1), 1-6.
Abbaspour, K. C., Yang, J., Maximov, I., Siber, R., Bogner, K., Mieleitner, J., ... & Srinivasan, R.
(2007). Modelling hydrology and water quality in the pre-alpine/alpine Thur watershed using
SWAT. Journal of hydrology, 333(2-4), 413-430.
Al-Busaidi, A. S., & Cookson, P. (2003). Salinity–pH relationships in calcareous soils. Journal of
Agricultural and Marine Sciences [JAMS], 8(1), 41-46.
Araújo, W. L., Nunes-Nesi, A., & Fernie, A. R. (2014). On the role of plant mitochondrial
metabolism and its impact on photosynthesis in both optimal and sub-optimal growth
conditions. Photosynthesis research, 119, 141-156.
Arce, M. I., Gómez, R., Suárez, M. L., & Vidal-Abarca, M. R. (2013). Denitrification rates and
controlling factors in two agriculturally influenced temporary Mediterranean saline streams.
Hydrobiologia, 700, 169-185.
Ardón, M., Morse, J. L., Colman, B. P., & Bernhardt, E. S. (2013). Drought‐induced saltwater
incursion leads to increased wetland nitrogen export. Global change biology, 19(10), 2976-
2985.
Aryal, B., Gurung, R., Camargo, A. F., Fongaro, G., Treichel, H., Mainali, B., ... & Puadel, S. R.
(2022). Nitrous oxide emission in altered nitrogen cycle and implications for climate change.
Environmental Pollution, 120272.
Chapter 2
87
Bayabil, H. K., Li, Y., Tong, Z., & Gao, B. (2021). Potential management practices of saltwater
intrusion impacts on soil health and water quality: a review. Journal of Water and Climate
Change, 12(5), 1327-1343.
Bennett, E., Carpenter, S., Gordon, L., Ramankutty, N., Balvanera, P., Campbell, B., ... &
Spierenburg, M. (2014). Toward a more resilient agriculture. Solutions, 5(5), 65-75.
Bernal, S., Hedin, L. O., Likens, G. E., Gerber, S., & Buso, D. C. (2012). Complex response of the
forest nitrogen cycle to climate change. Proceedings of the National Academy of Sciences,
109(9), 3406-3411.
Bernhardt, E. S., Palmer, M. A., Allan, J. D., Alexander, G., Barnas, K., Brooks, S., ... & Sudduth,
E. (2005). Synthesizing US river restoration efforts. Science, 308(5722), 636-637.
Bhardwaj, A. K., Nagaraja, M. S., Srivastava, S., Singh, A. K., & Arora, S. (2016). A framework
for adaptation to climate change effects in salt affected agricultural areas of Indo-Gangetic
region. Journal of Soil and Water Conservation, 15(1), 22-30.
Bieger, K., Arnold, J. G., Rathjens, H., White, M. J., Bosch, D. D., Allen, P. M., ... & Srinivasan,
R. (2017). Introduction to SWAT+, a completely restructured version of the soil and water
assessment tool. JAWRA Journal of the American Water Resources Association, 53(1), 115-
130.
Birgand, F., Appelboom, T. W., Chescheir, G. M., & Skaggs, R. W. (2011). Estimating nitrogen,
phosphorus, and carbon fluxes in forested and mixed-use watersheds of the lower coastal plain
of North Carolina: Uncertainties associated with infrequent sampling. Transactions of the
ASABE, 54(6), 2099-2110.
Chapter 2
88
Bond-Lamberty, B., Haddock, L. M., Pennington, S. C., Sezen, U. U., Shue, J., & Megonigal, J.
P. (2023). Salinity exposure affects lower-canopy specific leaf area of upland trees in a coastal
deciduous forest. Forest Ecology and Management, 548, 121404.
Bouwman, A. F., Bierkens, M. F. P., Griffioen, J., Hefting, M. M., Middelburg, J. J., Middelkoop,
H., & Slomp, C. P. (2013). Nutrient dynamics, transfer and retention along the aquatic
continuum from land to ocean: towards integration of ecological and biogeochemical models.
Biogeosciences, 10(1), 1-22.
Cazenave, A., & Cozannet, G. L. (2014). Sea level rise and its coastal impacts. Earth's Future,
2(2), 15-34.
Christensen, N. L., Bartuska, A. M., Brown, J. H., Carpenter, S., d'Antonio, C., Francis, R., ... &
Woodmansee, R. G. (1996). The report of the Ecological Society of America committee on the
scientific basis for ecosystem management. Ecological applications, 6(3), 665-691.
Church, J. A., Godfrey, J. S., Jackett, D. R., & McDougall, T. J. (1991). A model of sea level rise
caused by ocean thermal expansion. Journal of Climate, 4(4), 438-456.
Cohen, I., Huang, Y., Chen, J., Benesty, J., Benesty, J., Chen, J., ... & Cohen, I. (2009). Pearson
correlation coefficient. Noise reduction in speech processing, 1-4.
Conan, C., Bouraoui, F., Turpin, N., de Marsily, G., & Bidoglio, G. (2003). Modeling flow and
nitrate fate at catchment scale in Brittany (France). Journal of Environmental Quality, 32(6),
2026-2032.
Conrad, P. E., & Marinos, R. E. (2024). Nitrogen availability and denitrification in urban
agriculture and regreened vacant lots. Urban Ecosystems, 1-12.
Chapter 2
89
Craswell, E. (2021). Fertilizers and nitrate pollution of surface and ground water: an increasingly
pervasive global problem. SN Applied Sciences, 3(4), 518.
Cui, J., Jin, Z., Wang, Y., Gao, S., Fu, Z., Yang, Y., & Wang, Y. (2021). Mechanism of
eutrophication process during algal decomposition at the water/sediment interface. Journal of
Cleaner Production, 309, 127175.
Cui, Z., Huang, J., Gao, J., & Han, J. (2022). Characterizing the impacts of macrophyte-dominated
ponds on nitrogen sources and sinks by coupling multiscale models. Science of The Total
Environment, 811, 152208.
Donmez, C., Sari, O., Berberoglu, S., Cilek, A., Satir, O., & Volk, M. (2020). Improving the
applicability of the SWAT model to simulate flow and nitrate dynamics in a flat data-scarce
agricultural region in the Mediterranean. Water, 12(12), 3479.
Donner, S. D., & Kucharik, C. J. (2003). Evaluating the impacts of land management and climate
variability on crop production and nitrate export across the Upper Mississippi Basin. Global
Biogeochemical Cycles, 17(3).
Etesami, H., & Adl, S. M. (2020). Can interaction between silicon and non–rhizobial bacteria help
in improving nodulation and nitrogen fixation in salinity–stressed legumes? A review.
Rhizosphere, 15, 100229.
Etheridge, Randall J., Lepistö, A., Granlund, K., Rankinen, K., Birgand, F., & Burchell, M. R.
(2014). Reducing uncertainty in the calibration and validation of the INCA-N model by using
soft data. Hydrology Research, 45(1), 73-88.
Chapter 2
90
Fixen, P. E., & West, F. B. (2002). Nitrogen fertilizers: meeting contemporary challenges. Ambio:
a journal of the human environment, 31(2), 169-176.
Galloway, J. N. (1998). The global nitrogen cycle: changes and consequences. Environmental
pollution, 102(1), 15-24.
Galloway, J. N., Leach, A. M., Bleeker, A., & Erisman, J. W. (2013). A chronology of human
understanding of the nitrogen cycle. Philosophical Transactions of the Royal Society B:
Biological Sciences, 368(1621), 20130120.
Gerloff, G. C., & Krombholz, P. H. (1966). Tissue analysis as a measure of nutrient availability
for the growth of angiosperm aquatic plants 1. Limnology and Oceanography, 11(4), 529-537.
Gibson, N., McNulty, S., Miller, C., Gavazzi, M., Worley, E., Keesee, D., & Hollinger, D. (2021).
Identification, mitigation, and adaptation to salinization on working lands in the US Southeast.
Forest Service, US Department of Agriculture, Southern Research Station.
Gupta, A., Gupta, R., & Singh, R. L. (2017). Microbes and environment. Principles and
applications of environmental biotechnology for a sustainable future, 43-84.
Hamilton, D. P., Salmaso, N., & Paerl, H. W. (2016). Mitigating harmful cyanobacterial blooms:
strategies for control of nitrogen and phosphorus loads. Aquatic Ecology, 50, 351-366.
Hamonts, K., Clough, T. J., Stewart, A., Clinton, P. W., Richardson, A. E., Wakelin, S. A., ... &
Condron, L. M. (2013). Effect of nitrogen and waterlogging on denitrifier gene abundance,
community structure and activity in the rhizosphere of wheat. FEMS Microbiology Ecology,
83(3), 568-584.
Hart, M. H. (1978). The evolution of the atmosphere of the Earth. Icarus, 33(1), 23-39.
Chapter 2
91
Hattermann, F. F., Krysanova, V., Habeck, A., & Bronstert, A. (2006). Integrating wetlands and
riparian zones in river basin modelling. Ecological modelling, 199(4), 379-392.
Hay, W. W., & Leslie, M. A. (1990). Could possible changes in global groundwater reservoir cause
eustatic sea-level fluctuations. Sea-level change, 161-170.
He, W., Yang, J. Y., Qian, B., Drury, C. F., Hoogenboom, G., He, P., ... & Zhou, W. (2018).
Climate change impacts on crop yield, soil water balance and nitrate leaching in the semiarid
and humid regions of Canada. PloS one, 13(11), e0207370.
Helmers, M. J., Abendroth, L., Reinhart, B., Chighladze, G., Pease, L., Bowling, L., ... & Strock,
J. (2022). Impact of controlled drainage on subsurface drain flow and nitrate load: A synthesis
of studies across the US Midwest and Southeast. Agricultural Water Management, 259,
107265.
Herbert, E. R., Boon, P., Burgin, A. J., Neubauer, S. C., Franklin, R. B., Ardón, M., ... & Gell, P.
(2015). A global perspective on wetland salinization: ecological consequences of a growing
threat to freshwater wetlands. Ecosphere, 6(10), 1-43.
Hofstra, N., & Bouwman, A. F. (2005). Denitrification in agricultural soils: summarizing
published data and estimating global annual rates. Nutrient Cycling in Agroecosystems, 72,
267-278.
Jewell, W. J., & McCarty, P. L. (1971). Aerobic decomposition of algae. Environmental Science
& Technology, 5(10), 1023-1031.
Chapter 2
92
Karl, T. R., Melillo, J. M., & Peterson, T. C. (2009). Global climate change impacts in the United
States: a state of knowledge report from the US Global Change Research Program. Cambridge
University Press.
Kirova, E., & Kocheva, K. (2021). Physiological effects of salinity on nitrogen fixation in
legumes–a review. Journal of Plant Nutrition, 44(17), 2653-2662.
Kirwan, M. L., & Megonigal, J. P. (2013). Tidal wetland stability in the face of human impacts
and sea-level rise. Nature, 504(7478), 53-60.
Knighton, A. D., Mills, K., & Woodroffe, C. D. (1991). Tidal-creek extension and saltwater
intrusion in northern Australia. Geology, 19(8), 831-834.
Kuenzler, E. J., Stanley, D. W., & Koenings, J. R. (1979). Nutrient kinetics of phytoplankton in
the Pamlico River, North Carolina. Water Resources Research Institute of the University of
North Carolina.
Mathewson, P. D., Evans, S., Byrnes, T., Joos, A., & Naidenko, O. V. (2020). Health and economic
impact of nitrate pollution in drinking water: a Wisconsin case study. Environmental
Monitoring and Assessment, 192(11), 724.
Mazhar, S., Pellegrini, E., Contin, M., Bravo, C., & De Nobili, M. (2022). Impacts of salinization
caused by sea level rise on the biological processes of coastal soils-A review. Frontiers in
Environmental Science, 10, 909415.
Mazhar, S., Pellegrini, E., Contin, M., Bravo, C., & De Nobili, M. (2022). Impacts of salinization
caused by sea level rise on the biological processes of coastal soils-A review. Frontiers in
Environmental Science, 10, 909415.
Chapter 2
93
Mazhar, S., Pellegrini, E., Contin, M., Bravo, C., & De Nobili, M. (2022). Impacts of salinization
caused by sea level rise on the biological processes of coastal soils-A review. Frontiers in
Environmental Science, 10, 909415.
Melino, V., & Tester, M. (2023). Salt-tolerant crops: time to deliver. Annual Review of Plant
Biology, 74, 671-696.
Meunier, C. L., Gundale, M. J., Sánchez, I. S., & Liess, A. (2016). Impact of nitrogen deposition
on forest and lake food webs in nitrogen‐limited environments. Global Change Biology, 22(1),
164-179.
Mohd, F. A., Maulud, K. A., Karim, O. A., Begum, R. A., Awang, N. A., Hamid, M. A., ... & Abd
Razak, A. H. (2018, June). Assessment of coastal inundation of low lying areas due to sea level
rise. In IOP Conference Series: Earth and Environmental Science (Vol. 169, No. 1, p. 012046).
IOP Publishing.
Moraglia, G., Brattich, E., & Carbone, G. (2022). Precipitation trends in North and South Carolina,
USA. Journal of Hydrology: Regional Studies, 44, 101201.
Munksgaard, N. C., Hutley, L. B., Metcalfe, K. N., Padovan, A. C., Palmer, C., & Gibb, K. S.
(2019). Environmental challenges in a near-pristine mangrove estuary facing rapid urban and
industrial development: Darwin Harbour, Northern Australia. Regional studies in marine
science, 25, 100438.
NADP National Atmospheric Deposition Program. (2023). Retrieved December, 2023, from
https://nadp.slh.wisc.edu/
Chapter 2
94
National Oceanic and Atmospheric Administration (NOAA). (2022). Sea level rise technical
report. Retrieved from https://oceanservice.noaa.gov/hazards/sealevelrise/sealevelrise-tech-
report.html#step1
NCAGR North Carolina Department of Agriculture and Consumer Services. (2022). Tar-Pamlico
Watershed Initiatives. North Carolina Department of Agriculture and Consumer Services.
https://www.ncagr.gov/divisions/soil-water-conservation/programs-initiatives/watershed-
initiatives/tar-pamlico#Contact-4304
NCDEQ North Carolina Department of Environmental Quality. (2014). Tar-Pamlico Basin Plan.
North Carolina Department of Environmental Quality.
https://www.deq.nc.gov/about/divisions/water-resources/water-planning/basin-
planning/river-basin-plans/tar-pamlico#2014Tar-PamlicoBasinPlan-4040
Nelson, J. L., & Zavaleta, E. S. (2012). Salt marsh as a coastal filter for the oceans: changes in
function with experimental increases in nitrogen loading and sea-level rise.
Oh, J., & Sankarasubramanian, A. (2012). Interannual hydroclimatic variability and its influence
on winter nutrient loadings over the Southeast United States. Hydrology and Earth System
Sciences, 16(7), 2285-2298.
Okon, O. G. (2019). Effect of salinity on physiological processes in plants. Microorganisms in
saline environments: strategies and functions, 237-262.
Oliver-Smith, A. (2009). Sea level rise and the vulnerability of coastal peoples: responding to the
local challenges of global climate change in the 21st century. UNU-EHS.
Chapter 2
95
O’Brien, A., Townsend, K., Hale, R., Sharley, D., & Pettigrove, V. (2016). How is ecosystem
health defined and measured? A critical review of freshwater and estuarine studies. Ecological
Indicators, 69, 722-729.
Postgate, J. (1998). Nitrogen fixation. Cambridge University Press.
Preston, S. D., V. J. Bierman Jr., and S. E. Silliman. 1989. An evaluation of methods for the
estimation of tributary mass loads.Water Resour. Res. 25(6): 1379‐1389.
Pringle, C. M. (2001). Hydrologic connectivity and the management of biological reserves: a
global perspective. Ecological Applications, 11(4), 981-998.
Qian, J. H., Doran, J. W., Weier, K. L., Mosier, A. R., Peterson, T. A., & Power, J. F. (1997). Soil
denitrification and nitrous oxide losses under corn irrigated with high‐nitrate groundwater
(Vol. 26, No. 2, pp. 348-360). American Society of Agronomy, Crop Science Society of
America, and Soil Science Society of America.
Qian, L., Wang, F., Cao, W., Ding, S., & Cao, W. (2023). Ecological health assessment and
sustainability prediction in coastal area: A case study in Xiamen Bay, China. Ecological
Indicators, 148, 110047.
Rabalais, N. N. (2002). Nitrogen in aquatic ecosystems. AMBIO: a Journal of the Human
Environment, 31(2), 102-112.
Rasiah, V., Armour, J. D., & Cogle, A. L. (2005). Assessment of variables controlling nitrate
dynamics in groundwater: Is it a threat to surface aquatic ecosystems?. Marine pollution
bulletin, 51(1-4), 60-69.
Chapter 2
96
Rhoades, J. D., & Corwin, D. L. (1990). Soil electrical conductivity: effects of soil properties and
application to soil salinity appraisal. Communications in soil science and plant analysis, 21(11-
12), 837-860.
Rotzoll, K., & Fletcher, C. H. (2013). Assessment of groundwater inundation as a consequence of
sea-level rise. Nature Climate Change, 3(5), 477-481.
Safdar, H., Amin, A., Shafiq, Y., Ali, A., Yasin, R., Shoukat, A., ... & Sarwar, M. I. (2019). A
review: Impact of salinity on plant growth. Nat. Sci, 17(1), 34-40.
Saravanan, S., Singh, L., Sathiyamurthi, S., Sivakumar, V., Velusamy, S., & Shanmugamoorthy,
M. (2023). Predicting phosphorus and nitrate loads by using SWAT model in Vamanapuram
River Basin, Kerala, India. Environmental Monitoring and Assessment, 195(1), 186.
Sayemuzzaman, M., & Jha, M. K. (2014). Seasonal and annual precipitation time series trend
analysis in North Carolina, United States. Atmospheric Research, 137, 183-194.
Schilling, K. E., & Wolter, C. F. (2009). Modeling nitrate-nitrogen load reduction strategies for
the Des Moines River, Iowa using SWAT. Environmental management, 44(4), 671-682.
Schuerz, C., Hsieh, N.-H., & Willem. (2022). chrisschuerz/SWATplusR: SWATplusR 0.6
(Version 0.6) . Zenodo. https://doi.org/10.5281/zenodo.6517027
Seibel, B. A. (2011). Critical oxygen levels and metabolic suppression in oceanic oxygen
minimum zones. Journal of Experimental Biology, 214(2), 326-336.
Sekhon, G. S. (1995). Fertilizer-N use efficiency and nitrate pollution of groundwater in
developing countries. Journal of Contaminant Hydrology, 20(3-4), 167-184.
Chapter 2
97
Sherman, K. (1994). Sustainability, biomass yields, and health of coastal ecosystems: an ecological
perspective. Marine ecology progress series. Oldendorf, 112(3), 277-301.
Spivak, A. C., Sanderman, J., Bowen, J. L., Canuel, E. A., & Hopkinson, C. S. (2019). Global-
change controls on soil-carbon accumulation and loss in coastal vegetated ecosystems. Nature
Geoscience, 12(9), 685-692.
Stuart, M. E., Gooddy, D. C., Bloomfield, J. P., & Williams, A. T. (2011). A review of the impact
of climate change on future nitrate concentrations in groundwater of the UK. Science of the
Total Environment, 409(15), 2859-2873.
SWAT+ IO Document. (2020). Input/output file documentation, version 2016 modified on
November 16, 2020 according to REV 60.5. Retrieved from https://swatplus.gitbook.io/io-
docs
Temperton, V. M., Mwangi, P. N., Scherer-Lorenzen, M., Schmid, B., & Buchmann, N. (2007).
Positive interactions between nitrogen-fixing legumes and four different neighbouring species
in a biodiversity experiment. Oecologia, 151, 190-205.
Tian, L., Zhang, Y., & Zhu, J. (2014). Decreased surface albedo driven by denser vegetation on
the Tibetan Plateau. Environmental Research Letters, 9(10), 104001.
Upadhyay, P., Linhoss, A., Kelble, C., Ashby, S., Murphy, N., & Parajuli, P. B. (2022).
Applications of the SWAT model for coastal watersheds: review and recommendations.
Journal of the ASABE, 65(2), 453-469.
Ury, E. A., Yang, X., Wright, J. P., & Bernhardt, E. S. (2021). Rapid deforestation of a coastal
landscape driven by sea‐level rise and extreme events. Ecological applications, 31(5), e02339.
Chapter 2
98
Van Wijnen, H. J., & Bakker, J. P. (1999). Nitrogen and phosphorus limitation in a coastal barrier
salt marsh: the implications for vegetation succession. Journal of Ecology, 87(2), 265-272.
Vilar-Sanz, A., Puig, S., Garcia-Lledo, A., Trias, R., Balaguer, M. D., Colprim, J., & Baneras, L.
(2013). Denitrifying bacterial communities affect current production and nitrous oxide
accumulation in a microbial fuel cell. PLoS One, 8(5), e63460.
Vilmin, L., Mogollón, J. M., Beusen, A. H., & Bouwman, A. F. (2018). Forms and subannual
variability of nitrogen and phosphorus loading to global river networks over the 20th century.
Global and Planetary Change, 163, 67-85.
Vitousek, P. M., Aber, J. D., Howarth, R. W., Likens, G. E., Matson, P. A., Schindler, D. W., ...
& Tilman, D. G. (1997). Human alteration of the global nitrogen cycle: sources and
consequences. Ecological applications, 7(3), 737-750.
Voss, B. M., Peucker-Ehrenbrink, B., Eglinton, T. I., Spencer, R. G., Bulygina, E., Galy, V., ... &
Luymes, R. (2015). Seasonal hydrology drives rapid shifts in the flux and composition of
dissolved and particulate organic carbon and major and trace ions in the Fraser River, Canada.
Biogeosciences, 12(19), 5597-5618.
Wang, H., Steyer, G. D., Couvillion, B. R., Beck, H. J., Rybczyk, J. M., Rivera‐Monroy, V. H., ...
& Visser, J. M. (2017). Predicting landscape effects of Mississippi River diversions on soil
organic carbon sequestration. Ecosphere, 8(11), e01984.
Wang, J., Yin, H., & Chung, F. (2011). Isolated and integrated effects of sea level rise, seasonal
runoff shifts, and annual runoff volume on California’s largest water supply. Journal of
Hydrology, 405(1-2), 83-92.
Chapter 2
99
Wurtsbaugh, W. A., Paerl, H. W., & Dodds, W. K. (2019). Nutrients, eutrophication and harmful
algal blooms along the freshwater to marine continuum. Wiley Interdisciplinary Reviews:
Water, 6(5), e1373.
Zhang, Y., Xia, J., Chen, J., & Zhang, M. (2011). Water quantity and quality optimization
modeling of dams operation based on SWAT in Wenyu River Catchment, China.
Environmental Monitoring and Assessment, 173, 409-430.
Zhu, X., Zhang, W., Chen, H., & Mo, J. (2015). Impacts of nitrogen deposition on soil nitrogen
cycle in forest ecosystems: A review. Acta Ecologica Sinica, 35(3), 35-43.
Chapter 3a: Framework for Stakeholder-Driven Socio-Hydrological Modeling: Conceptual
Foundations for Policy Development and Evaluation to Improve Ecosystem Health
Mahesh R Tapasa*, Randall Etheridgeb, Gregory Howardc, Matthew Maird
a Integrated Coastal Program, East Carolina University, Greenville, NC 27858, USA
b Department of Engineering, Center for Sustainable Energy and Environmental Engineering,
East Carolina University, Greenville, NC 27858, USA
c Department of Economics, East Carolina University, Greenville, NC 27858, USA
d Department of Economics, Appalachian State University, Boone, NC 28608
Abstract
In this paper, we propose a concise framework that integrates stakeholder insights, specifically
from farmers, into hydrological modeling (Soil and Water Assessment Tool Plus, SWAT+), to
enhance the efficacy of potential watershed management policies. By incorporating farmers'
behavioral responses to policy changes within the SWAT+ model, we bridge the gap between
econometric and hydrological modeling. This interdisciplinary approach captures the dynamic
interactions between stakeholders and ecosystem health, offering a novel perspective on
environmental management. Our methodology includes surveys to gauge farmers' reactions to
policy changes and quantify the adoption of agricultural best management practices, forecasting
the impact of policy changes on land management and subsequent impacts on ecosystem health.
This integration facilitates a more realistic representation of socio-environmental systems,
underscoring the importance of aligning policy interventions with the practical realities faced by
those directly affecting and interacting with the ecosystems. Through this framework, we propose
a pathway towards more effective and sustainable environmental policies, emphasizing the critical
Chapter 3a
101
role of incorporating stakeholder perspectives into model development and decision-making
processes.
Keywords
Socio-hydrological modeling, Stakeholder engagement, Behavioral econometrics, Policy
development, Ecosystem health
1 Introduction
1.1 Context and challenge
The evolving landscape of environmental challenges underscores a pressing need to incorporate
social dynamics into hydrological modeling (Archfield et al., 2015). Traditional models, primarily
focused on physical and biological processes (Vereecken et al., 2016), often oversimplify or
overlook the complex interplay between human activities and natural systems (Blair & Buytaert,
2016). This oversight is particularly consequential in coastal ecosystems, where the confluence of
land and water creates unique environmental settings (Garcia 2003; Kelly et al., 2018). These areas
are not only hotspots for biodiversity but also for human settlement (Kong et al., 2021) and
economic activities (Marchese 2015), making them especially vulnerable to the impacts of land
use changes (Li et al., 2007), pollution (Blanchard & Lerch, 2000), and climate change (Schulze,
1997).
Hydrological modelers attempt to simulate the complex relationship between water systems
and land management (McGrane, 2016), where human activities like agriculture (Curk & Glavan,
2021) and urban development (Salvadore et al., 2015) impact water quality (Keller et al., 2023)
and ecosystem health, and vice versa (Qiu et al., 2021). Integrating social dynamics into
Chapter 3a
102
hydrological models without losing scientific accuracy or efficiency poses a significant challenge,
as factors like human decision-making, cultural norms, and socioeconomic conditions could
introduce variability and unpredictability in model simulations (Xu et al., 2018). Ecosystems,
affected by human actions and natural factors (Zhao et al., 2023), should use a new approach to
modeling that combines social sciences (Forrester et al., 2014), economics (Bockstael et al., 1995),
and traditional hydrology (Walters, 1997) for models that accurately reflect socio-environmental
interactions (Lawyer et al., 2023; Tanim et al., 2022).
Incorporating social dynamics into hydrological models improves their relevance for
policymaking (Blair & Buytaert, 2016) providing insights into the impact of interventions on both
ecosystems (Elsawah et al., 2017) and communities (Givens et al., 2018). This holistic approach
helps anticipate the outcomes of management actions (Mostert, 2018), leading to better
environmental stewardship (Sivapalan et al., 2012) and sustainable management of coastal land
and ecosystems (Kumar et al., 2020). Thus, integrating social factors into modeling is crucial for
sustainable land management (De Jonge, 2012).
1.2 Objective
This work introduces a conceptual framework for integrating socio-economic aspects into
hydrological modeling, aiming to improve model accuracy and policy relevance. By combining
social and hydrological sciences, the framework enhances sustainable land management by
considering human impact on downstream systems, increasing the likelihood that adopted
strategies are environmentally and socio-economically sustainable.
1.3 Significance
Integrating socio-hydrological elements into hydrological models can enhance both theoretical
understanding and practical application, leading to more accurate and policy-relevant outcomes
Chapter 3a
103
(Mostert, 2018; Troy et al., 2015). Such an approach recognizes the interconnectedness of water
systems with human actions, improving model realism by accounting for human variability
(Schlueter et al., 2012). Practically, successful integration could make water management policies
more effective by addressing the root causes of issues (Jønch-Clausen, 2004), predicting human
responses (Srinivasan et al., 2017), and supporting sustainable practices (Sivapalan et al., 2014).
This approach has the potential to promote scientifically robust (Vogel et al., 2015), socially
equitable (Mirza & Mustafa, 2016), and economically viable strategies (Sivapalan, 2015),
improving stakeholder engagement and policy success (Xu et al., 2018).
2 Theoretical Background of Socio-Hydrological Integration
2.1 Socio-hydrology
Socio-hydrology studies the interaction between human societies and water systems, examining
how human activities affect hydrology and how water changes influence human life (Sivapalan et
al., 2012). It views water as essential for ecosystems, agriculture, and economies, while examining
human impacts on the hydrological cycle (Abbott et al., 2019). As global anthropogenic challenges
like population growth and climate change intensify water issues, socio-hydrology's holistic
approach combines social and hydrological sciences to foster sustainable water management,
guiding the creation of policies that meet both human and environmental needs (Levy et al., 2016;
Pande & Sivapalan, 2017).
Integrating socio-economic perspectives with hydrology is key to solving complex water issues
beyond just technical means, addressing water conflicts, ecosystem degradation, and community
resilience to floods and droughts (Everard, 2019; Pande et al., 2020). Socio-hydrology identifies
Chapter 3a
104
change drivers and promotes adaptive, inclusive water management (Xu et al., 2018). This
emphasizes the need for interdisciplinary research and collaborative policymaking to address
challenges at the intersection of water and society (Krueger et al., 2016).
2.2 Challenges of integration
Combining social science with hydrological modeling faces challenges due to different approaches
and scales (Xu et al., 2018). Modeling complex human behavior and integrating it with
hydrological process models is difficult because of the inherent unpredictability and non-linearity
of social systems. (Blair & Buytaert, 2016). Developing new tools that accommodate both
(quantitative and qualitative) social data and quantitative hydrological data is challenging, given
the varying temporal and spatial scales of social and hydrological processes (Fischer et al., 2021).
The integration of social variables into hydrological models is hindered by the scarcity,
inconsistency, or absence of social data, posing significant challenges for accurate modeling
(Brunner et al., 2021; Muñoz-Carpena et al., 2023). Furthermore, effective interdisciplinary
collaboration is difficult due to differences in terminology and concepts between fields (Moirano
et al., 2020). Validating socio-hydrological models also presents difficulties, as traditional
hydrological validation methods do not suit models with social dynamics, necessitating new
approaches to ensure model reliability (Iwanaga et al., 2020). Overcoming these challenges
demands innovative and interdisciplinary efforts.
3 Econometric Modeling in Socio-Hydrological Contexts
3.1 Basics
Econometric modeling, the use of statistical methods to analyze economic data (broadly defined),
is a key component of the socio-hydrological framework, allowing researchers to understand how
Chapter 3a
105
economic policies affect water-related human behaviors (Alamanos & Zeng, 2021; Bertassello et
al., 2021). Economic data is defined as data on individuals’ resource allocation decisions, and as
such involves a wide array of financial and non-financial human decision-making. Econometric
modeling quantifies the impact of policies on resource use and management, aiding in forecasting
policy outcomes (McColl & Aggett, 2007; Daneshi et al., 2021). By examining the effects of
economic incentives on water conservation practices and land use changes, econometric modeling
helps create realistic models that blend human decision-making into hydrological simulations
(Castro & Lechthaler, 2022; Ovando & Brouwer, 2019). This integrated approach is vital for
devising effective land management strategies that account for both environmental processes and
socio-economic factors.
3.2 Advantages
Incorporating econometric insights into hydrological models enhances their utility by improving
human behavior predictions, enabling data-driven policy impact assessments, and fostering an
interdisciplinary approach to land management (Dinar, 2024; Jarvis et al., 2013). This integration
allows for targeted policy design by explaining the socio-economic drivers of land use and
evaluating various policy effects on both human behavior and hydrological systems (Conrad, 2016;
Ovando & Brouwer, 2019). Such integration bridges the gap between social science and hydrology,
leading to a comprehensive understanding of the interaction between ecological processes and
socio-economic factors, and thus promoting holistic solutions to water challenges (Anderson et al.,
2019; Di Baldassarre et al., 2019). In human-dominated watersheds, a resource management
policy’s impact on the environment is heavily mediated through the ways in which the policy alters
human action. By adding a layer of social realism, econometric modeling turns hydrological
Chapter 3a
106
models into more dynamic and relevant tools for addressing water management issues (Jarvis et
al., 2013; Liu et al., 2019).
4 Framework for Integrating Engineering and Economic Models
4.1 Methodological overview
Payment for ecosystem services (PES) programs are a recent policy approach designed to
incentivize the adoption of agricultural best management practices (BMPs). PES programs offer
financial compensation to landowners for the voluntary implementation of BMPs and other
ecosystem services (Ma et al., 2012). There have been several PES programs implemented by the
USDA designed to reduce agricultural nutrient pollution, such as the Conservation Reserve
Enhancement Program (CREP), Conservation Stewardship Program (CSP), and Environmental
Quality Incentives Program (EQIP) (Howard et al., 2023).
Survey data is often employed to study how stakeholders (in this example, farmers) would
respond to a diverse set of PES alternatives. Discrete choice experiments can be included in
surveys for this purpose. Discrete choice experiments require respondents to choose between two
or more hypothetical alternatives. Each alternative differs along a variety of attributes. For
example, a farmer may be asked to choose between two hypothetical conservation contracts, which
present different requirements for the adoption of BMPs and varying levels of compensation, and
a third option of rejecting both contracts. This methodology can be applied to a wide range of
contexts including other hypothetical policy scenarios.
By thoughtfully and methodically varying the alternatives presented in a discrete choice
experiment, the choice data can be used to tease out how each element of a choice (in discrete
Chapter 3a
107
choice modeling parlance, each attribute) impacts a decision maker’s willingness to choose a given
alternative. This process, typically done with a logistic regression model, has been replicated
several times in the socio-hydrological context of land-owner preferences for adopting BMPs
(Beharry-Borg et al., 2013; Howard et al., 2023; Kaczan et al., 2013; Raes et al., 2017).
By far, the most common application of econometric modeling in the socio-hydrological
context is through the random utility maximization (RUM) model (Manski, 1977). The RUM
framework posits that the utility a decision maker receives from an alternative is a function of a
set of observable (to the researcher) characteristics and an unobserved random component. This
framework can account for heterogeneity in preferences for observable factors, such as
demographics and attitudes, and is particularly useful when the survey sample is diverse. The
random utility model utilizes discrete choice experiment data and maximum likelihood (or
maximum simulated likelihood) estimation to generate marginal utility estimates for each element
of the observable characteristics of a choice. While utility is an abstract concept of well-being or
satisfaction, these estimates are valuable for calculating probabilities that an alternative will be
selected as well as willingness-to-accept (WTA) values. In the context of PES contracts, the
willingness to accept value for a given alternative represents the minimum compensation necessary
to induce the decision maker to agree to the PES contract.
Assuming the sample that is used to parameterize one’s econometric model is representative
of the target population, it is a relatively straightforward exercise to project the number of people
who would adopt or reject a hypothetical policy. In the context of environmental subsidies or PES
contracts, the people who adopt the policy are those whose estimated WTA values are below the
payment offered by a policy. This method is demonstrated by Howard et al., (2023), who examine
participation in agricultural conservation programs in Iowa.
Chapter 3a
108
The theoretical integration of engineering models (Figure 3a.1), such as SWAT+ (Soil and
Water Assessment Tool Plus), with econometric models presents a methodological advancement
in socio-hydrological modeling. This integration is designed to capture the complexity of
interactions between human behaviors, policy interventions, and hydrological systems. The
framework proposed here operates on a modular basis, where each model—engineering and
econometric—functions as an independent module that interacts through well-defined interfaces.
Figure 3a.1 Framework for Integrating SWAT+ and Econometric Models in Socio-Hydrological
Studies
Chapter 3a
109
At its core, the SWAT+ model provides a detailed simulation of hydrological processes
(Wagner et al., 2022), including water flow, sediment transport, and nutrient cycling across a
watershed (Van et al., 2021). While highly effective at these goals, SWAT+ provides little to no
insight into how land use and land management practices will change in response to policy stimuli.
Econometric models quantify the influence of socio-economic factors on human decisions,
including how changes in policy or environmental conditions might affect land use, agricultural
practices, and water usage (Zessner et al., 2017). The integration process of these two modeling
frameworks involves aligning the outputs of econometric models—essentially predictions of
human behavior and land use change—with the input requirements of SWAT+, thereby enabling
the hydrological model to simulate the effects of these socio-economic changes on water systems.
This methodological integration is facilitated by a data exchange layer, which translates the
econometric model's outputs into parameters and variables that are understandable and usable by
SWAT+. For instance, changes in crop planting decisions, derived from the econometric model,
are input into SWAT+ as changes in land cover, which then influences runoff, evapotranspiration,
and nutrient loading in the simulation. Similarly, policy changes affecting water usage or
conservation practices are reflected in SWAT+ through adjustments to irrigation practices or the
implementation of BMPs. A key advantage of this modular approach is its flexibility; the
framework can be adapted or expanded to include additional models or data sources as new
information becomes available or as research objectives evolve. Moreover, this integration allows
for iterative refinement, where outputs from the combined model can inform further adjustments
to the econometric assumptions or SWAT+ parameters, enhancing the overall accuracy and
relevance of the socio-hydrological analysis. This methodological overview underscores the
framework's potential to provide comprehensive insights into the reciprocal influences of human
Chapter 3a
110
activities and natural systems, thereby supporting more informed and effective resource
management and policy development.
4.2 Potential applications
The integrated framework merging SWAT+ engineering models with econometric models
facilitates addressing complex environmental and policy issues through a multidisciplinary lens.
This approach enables detailed analysis of the interaction between human activities and water
systems, supporting informed water resource management decisions.
Impact Assessment of Agro-environmental Policies: Evaluates how agricultural policy changes
affect farming practices and their impact on water quality and flow.
Water Resource Management Under Climate Change: Assesses the effects of climate change on
water availability and demand, supporting the development of adaptive strategies.
Urbanization and Its Hydrological Impacts: Examines the hydrological effects of urban expansion,
aiding in flood risk mitigation and sustainable urban development.
Ecosystem Services Valuation: Quantifies the economic benefits of ecosystem services related to
water systems, guiding investments in conservation.
Stakeholder Engagement in Water Governance: Simulates participatory governance models to
explore consensus-building and collaborative management approaches.
Evaluation of Water Conservation Measures: Analyzes individual and societal responses to water
conservation policies, essential for managing water scarcity.
Chapter 3a
111
4.3 Broader Implications
This integrated modeling framework significantly impacts sustainable coastal management and
climate adaptation strategies. It can partially evaluate the impacts of saltwater intrusion on farmers’
land use and land management. For climate adaptation, it assesses water management strategies
under future climate scenarios, considering both environmental sustainability and social equity.
By integrating engineering and economic models, this approach enhances policy-making, land use
planning, and ecosystem management (Boulanger & Bréchet, 2005). By offering a holistic view
of human-nature interactions to inform adaptive, this framework can better inform strategies for
environmental change and socio-economic development.
5 Conclusion and Future Directions
In summary, our theoretical exploration into Stakeholder-Driven Socio-Hydrological Modeling
illuminates the critical importance and vast potential of integrating socio-hydrological models for
enhancing land management and policy-making. This integration bridges the gap between
hydrological science and social dynamics, thereby enabling policies that are more attuned to the
complexities of human-environment interactions. Despite its promise, this approach faces
theoretical and practical limitations, including the challenge of accurately modeling human
behavior and the need for high-quality, socio-economic data. These hurdles underscore the
importance of interdisciplinary collaboration across hydrology, social sciences, and economics to
advance our methodological approaches. Future research should focus on refining these integration
techniques, improving data acquisition and validation strategies, and exploring innovative ways to
incorporate stakeholder insights into hydrological models.
LLM Statement
We utilized GPT-4.0 for grammar corrections and to enhance sentence flow. Following the AI's
corrections, all materials were carefully reviewed by the authors.
References
Abbott, B. W., Bishop, K., Zarnetske, J. P., Minaudo, C., Chapin III, F. S., Krause, S., ... & Pinay,
G. (2019). Human domination of the global water cycle absent from depictions and
perceptions. Nature Geoscience, 12(7), 533-540.
Alamanos, A., & Zeng, Q. (2021). Managing Scarce water resources for socially acceptable
solutions, through hydrological and econometric modeling. Central Asian Journal of Water
Research, 7(1), 84-101.
Anderson, E. P., Jackson, S., Tharme, R. E., Douglas, M., Flotemersch, J. E., Zwarteveen, M., ...
& Arthington, A. H. (2019). Understanding rivers and their social relations: A critical step to
advance environmental water management. Wiley Interdisciplinary Reviews: Water, 6(6),
e1381.
Archfield, S. A., Clark, M., Arheimer, B., Hay, L. E., McMillan, H., Kiang, J. E., ... & Over, T.
(2015). Accelerating advances in continental domain hydrologic modeling. Water Resources
Research, 51(12), 10078-10091.
Vereecken, H., Schnepf, A., Hopmans, J. W., Javaux, M., Or, D., Roose, T., ... & Young, I. M.
(2016). Modeling soil processes: Review, key challenges, and new perspectives. Vadose zone
journal, 15(5).
Chapter 3a
113
Babaeian, F., Delavar, M., Morid, S., & Jamshidi, S. (2023). Designing climate change dynamic
adaptive policy pathways for agricultural water management using a socio-hydrological
modeling approach. Journal of Hydrology, 627, 130398.
Beharry-Borg, N., Smart, J. C. R., Termansen, M., & Hubacek, K. (2013). Evaluating farmers’
likely participation in a payment programme for water quality protection in the UK uplands.
Regional Environmental Change, 13(3), 633–647. https://doi.org/10.1007/s10113-012-0282-9
Bertassello, L., Levy, M. C., & Müller, M. F. (2021). Sociohydrology, ecohydrology, and the space-
time dynamics of human-altered catchments. Hydrological Sciences Journal, 66(9), 1393-
1408.
Blair, P., & Buytaert, W. (2016). Socio-hydrological modelling: a review asking “why, what and
how?”. Hydrology and Earth System Sciences, 20(1), 443-478.
Blanchard, P. E., & Lerch, R. N. (2000). Watershed vulnerability to losses of agricultural
chemicals: Interactions of chemistry, hydrology, and land-use. Environmental science &
technology, 34(16), 3315-3322.
Bockstael, N., Costanza, R., Strand, I., Boynton, W., Bell, K., & Wainger, L. (1995). Ecological
economic modeling and valuation of ecosystems. Ecological economics, 14(2), 143-159.
Boulanger, P. M., & Bréchet, T. (2005). Models for policy-making in sustainable development:
The state of the art and perspectives for research. Ecological economics, 55(3), 337-350.
Brunner, M. I., Slater, L., Tallaksen, L. M., & Clark, M. (2021). Challenges in modeling and
predicting floods and droughts: A review. Wiley Interdisciplinary Reviews: Water, 8(3), e1520.
Chapter 3a
114
Castro, L. M., & Lechthaler, F. (2022). The contribution of bio-economic assessments to better
informed land-use decision making: An overview. Ecological Engineering, 174, 106449.
Conrad, S. A. (2016). The use of discrete choice experiments and a coupled socio-hydrological
model to inform water policymaking in the Okanagan region of British Columbia.
Curk, M., & Glavan, M. (2021). Perspectives of hydrologic modeling in agricultural research. In
Hydrology. Rijeka, Croatia: IntechOpen.
Daneshi, A., Brouwer, R., Najafinejad, A., Panahi, M., Zarandian, A., & Maghsood, F. F. (2021).
Modelling the impacts of climate and land use change on water security in a semi-arid forested
watershed using InVEST. Journal of Hydrology, 593, 125621.
De Jonge, V. N., Pinto, R., & Turner, R. K. (2012). Integrating ecological, economic and social
aspects to generate useful management information under the EU Directives'‘ecosystem
approach’. Ocean & Coastal Management, 68, 169-188.
Di Baldassarre, G., Sivapalan, M., Rusca, M., Cudennec, C., Garcia, M., Kreibich, H., ... &
Blöschl, G. (2019). Sociohydrology: scientific challenges in addressing the sustainable
development goals. Water Resources Research, 55(8), 6327-6355.
Dinar, A. (2024). Challenges to Water Resource Management: The Role of Economic and
Modeling Approaches. Water, 16(4), 610.
Elsawah, S., Pierce, S. A., Hamilton, S. H., Van Delden, H., Haase, D., Elmahdi, A., & Jakeman,
A. J. (2017). An overview of the system dynamics process for integrated modelling of socio-
ecological systems: Lessons on good modelling practice from five case studies. Environmental
Modelling & Software, 93, 127-145.
Chapter 3a
115
Everard, M. (2019). A socio-ecological framework supporting catchment-scale water resource
stewardship. Environmental science & policy, 91, 50-59.
Fischer, A., Miller, J. A., Nottingham, E., Wiederstein, T., Krueger, L. J., Perez-Quesada, G., ... &
Sanderson, M. R. (2021). A systematic review of spatial-temporal scale issues in
sociohydrology. Frontiers in Water, 3, 730169.
Forrester, J., Greaves, R., Noble, H., & Taylor, R. (2014). Modeling social‐ecological problems in
coastal ecosystems: A case study. Complexity, 19(6), 73-82.
Garcia, S. M. (2003). The ecosystem approach to fisheries: issues, terminology, principles,
institutional foundations, implementation and outlook (No. 443). Food & Agriculture Org..
Givens, J. E., Padowski, J., Guzman, C. D., Malek, K., Witinok-Huber, R., Cosens, B., ... & Adam,
J. (2018). Incorporating social system dynamics in the Columbia River Basin: Food-energy-
water resilience and sustainability modeling in the Yakima River Basin. Frontiers in
environmental science, 6, 104.
Howard, G., Zhang, W., Valcu‐Lisman, A., & Gassman, P. W. (2023). Evaluating the tradeoff
between cost effectiveness and participation in agricultural conservation programs. American
Journal of Agricultural Economics, ajae.12397. https://doi.org/10.1111/ajae.12397
Iwanaga, T., Partington, D., Ticehurst, J., Croke, B. F., & Jakeman, A. J. (2020). A socio-
environmental model for exploring sustainable water management futures: Participatory and
collaborative modelling in the Lower Campaspe catchment. Journal of Hydrology: Regional
Studies, 28, 100669.
Chapter 3a
116
Jarvis, D., Stoeckl, N., & Chaiechi, T. (2013). Applying econometric techniques to hydrological
problems in a large basin: Quantifying the rainfall–discharge relationship in the Burdekin,
Queensland, Australia. Journal of Hydrology, 496, 107-121.
Jarvis, D., Stoeckl, N., & Chaiechi, T. (2013). Applying econometric techniques to hydrological
problems in a large basin: Quantifying the rainfall–discharge relationship in the Burdekin,
Queensland, Australia. Journal of Hydrology, 496, 107-121.
Jønch-Clausen, T. (2004). Integrated water resources management (IWRM) and water efficiency
plans by 2005: Why, what and how. Why, what and how, 5-4.
Kaczan, D., Swallow, B. M., & Adamowicz, W. L. (Vic). (2013). Designing a payments for
ecosystem services (PES) program to reduce deforestation in Tanzania: An assessment of
payment approaches. Ecological Economics, 95, 20–30.
https://doi.org/10.1016/j.ecolecon.2013.07.011
Keller, A. A., Garner, K., Rao, N., Knipping, E., & Thomas, J. (2023). Hydrological models for
climate-based assessments at the watershed scale: A critical review of existing hydrologic and
water quality models. Science of The Total Environment, 867, 161209.
Kelly, C., Ellis, G., & Flannery, W. (2018). Conceptualising change in marine governance: learning
from transition management. Marine Policy, 95, 24-35.
Kong, X., Zhou, Z., & Jiao, L. (2021). Hotspots of land-use change in global biodiversity hotspots.
Resources, Conservation and Recycling, 174, 105770.
Chapter 3a
117
Krueger, T., Maynard, C., Carr, G., Bruns, A., Mueller, E. N., & Lane, S. (2016). A
transdisciplinary account of water research. Wiley Interdisciplinary Reviews: Water, 3(3), 369-
389.
Kumar, P., Avtar, R., Dasgupta, R., Johnson, B. A., Mukherjee, A., Ahsan, M. N., ... & Mishra, B.
K. (2020). Socio-hydrology: A key approach for adaptation to water scarcity and achieving
human well-being in large riverine islands. Progress in Disaster Science, 8, 100134.
Lawyer, C., An, L., & Goharian, E. (2023). A Review of Climate Adaptation Impacts and Strategies
in Coastal Communities: From Agent-Based Modeling towards a System of Systems
Approach. Water, 15(14), 2635.
Levy, M. C., Garcia, M., Blair, P., Chen, X., Gomes, S. L., Gower, D. B., ... & Zeng, R. (2016).
Wicked but worth it: student perspectives on socio-hydrology. Hydrol. Process, 30(9), 1467-
1472.
Li, K. Y., Coe, M. T., Ramankutty, N., & De Jong, R. (2007). Modeling the hydrological impact of
land-use change in West Africa. Journal of hydrology, 337(3-4), 258-268.
Liu, H., Gopalakrishnan, S., Browning, D., & Sivandran, G. (2019). Valuing water quality change
using a coupled economic-hydrological model. Ecological Economics, 161, 32-40.
Ma, S., Swinton, S. M., Lupi, F., & Jolejole-Foreman, C. (2012). Farmers’ willingness to
participate in payment-for-environmental-services programmes. Journal of Agricultural
Economics, 63(3), 604–626. https://doi.org/10.1111/j.1477-9552.2012.00358.x
Chapter 3a
118
Manski, C. (1977, July 1). The Structure of Random Utility Models.
https://www.proquest.com/openview/7acf07ef00e4d7b837b4de87994aed40/1?cbl=1818302&
pq-origsite=gscholar
Mao, F., Clark, J., Karpouzoglou, T., Dewulf, A., Buytaert, W., & Hannah, D. (2017). HESS
Opinions: A conceptual framework for assessing socio-hydrological resilience under change.
Hydrology and Earth System Sciences, 21(7), 3655-3670.
Marchese, C. (2015). Biodiversity hotspots: A shortcut for a more complicated concept. Global
Ecology and Conservation, 3, 297-309.
McColl, C., & Aggett, G. (2007). Land-use forecasting and hydrologic model integration for
improved land-use decision support. Journal of environmental management, 84(4), 494-512.
McGrane, S. J. (2016). Impacts of urbanisation on hydrological and water quality dynamics, and
urban water management: a review. Hydrological Sciences Journal, 61(13), 2295-2311.
Melsen, L. A., Vos, J., & Boelens, R. (2018). What is the role of the model in socio-hydrology?
Discussion of “Prediction in a socio-hydrological world”. Hydrological Sciences Journal,
63(9), 1435-1443.
Mijic, A., Liu, L., O’Keeffe, J., Dobson, B., & Chun, K. P. (2024). A meta-model of socio-
hydrological phenomena for sustainable water management. Nature Sustainability, 7(1), 7-14.
Mirza, M. U., & Mustafa, D. (2016). Access, equity and hazards: highlighting a socially just and
ecologically resilient perspective on water resources. Sustainable Development and Disaster
Risk Reduction, 143-159.
Chapter 3a
119
Moirano, R., Sánchez, M. A., & Štěpánek, L. (2020). Creative interdisciplinary collaboration: A
systematic literature review. Thinking Skills and Creativity, 35, 100626.
Mostert, E. (2018). An alternative approach for socio-hydrology: case study research. Hydrology
and Earth System Sciences, 22(1), 317-329.
Mostert, E. (2018). An alternative approach for socio-hydrology: case study research. Hydrology
and Earth System Sciences, 22(1), 317-329.
Muñoz-Carpena, R., Carmona-Cabrero, A., Yu, Z., Fox, G., & Batelaan, O. (2023). Convergence
of mechanistic modeling and artificial intelligence in hydrologic science and engineering.
PLOS Water, 2(8), e0000059.
Ovando, P., & Brouwer, R. (2019). A review of economic approaches modeling the complex
interactions between forest management and watershed services. Forest Policy and Economics,
100, 164-176.
Ovando, P., & Brouwer, R. (2019). A review of economic approaches modeling the complex
interactions between forest management and watershed services. Forest Policy and Economics,
100, 164-176.
Pande, S., & Sivapalan, M. (2017). Progress in socio‐hydrology: A meta‐analysis of challenges
and opportunities. Wiley Interdisciplinary Reviews: Water, 4(4), e1193.
Pande, S., Roobavannan, M., Kandasamy, J., Sivapalan, M., Hombing, D., Lyu, H., & Rietveld, L.
(2020). A socio-hydrological perspective on the economics of water resources development
and management. In Oxford Research Encyclopedia of Environmental Science.
Chapter 3a
120
Qiu, H., Qi, J., Lee, S., Moglen, G. E., McCarty, G. W., Chen, M., & Zhang, X. (2021). Effects of
temporal resolution of river routing on hydrologic modeling and aquatic ecosystem health
assessment with the SWAT model. Environmental Modelling & Software, 146, 105232.
Raes, L., Speelman, S., & Aguirre, N. (2017). Farmers’ Preferences for PES Contracts to Adopt
Silvopastoral Systems in Southern Ecuador, Revealed Through a Choice Experiment.
Environmental Management, 60(2), 200–215. https://doi.org/10.1007/s00267-017-0876-6
Salvadore, E., Bronders, J., & Batelaan, O. (2015). Hydrological modelling of urbanized
catchments: A review and future directions. Journal of hydrology, 529, 62-81.
Schlueter, M., Mcallister, R. R., Arlinghaus, R., Bunnefeld, N., Eisenack, K., Hoelker, F., ... &
Stöven, M. (2012). New horizons for managing the environment: A review of coupled social‐
ecological systems modeling. Natural Resource Modeling, 25(1), 219-272.
Schulze, R. E. (1997). Impacts of global climate change in a hydrologically vulnerable region:
challenges to South African hydrologists. Progress in physical geography, 21(1), 113-136.
Sivapalan, M. (2015). Debates—Perspectives on socio‐hydrology: Changing water systems and
the “tyranny of small problems”—Socio‐hydrology. Water Resources Research, 51(6), 4795-
4805.
Sivapalan, M., Konar, M., Srinivasan, V., Chhatre, A., Wutich, A., Scott, C. A., ... & Rodríguez‐
Iturbe, I. (2014). Socio‐hydrology: Use‐inspired water sustainability science for the
Anthropocene. Earth's Future, 2(4), 225-230.
Sivapalan, M., Savenije, H. H., & Blöschl, G. (2012). Socio-hydrology: A new science of people
and water. Hydrol. Process, 26(8), 1270-1276.
Chapter 3a
121
Sivapalan, M., Savenije, H. H., & Blöschl, G. (2012). Socio-hydrology: A new science of people
and water. Hydrol. Process, 26(8), 1270-1276.
Srinivasan, V., Sanderson, M., Garcia, M., Konar, M., Blöschl, G., & Sivapalan, M. (2017).
Prediction in a socio-hydrological world. Hydrological Sciences Journal, 62(3), 338-345.
Tanim, A. H., Goharian, E., & Moradkhani, H. (2022). Integrated socio-environmental
vulnerability assessment of coastal hazards using data-driven and multi-criteria analysis
approaches. Scientific Reports, 12(1), 11625.
Troy, T. J., Pavao‐Zuckerman, M., & Evans, T. P. (2015). Debates—Perspectives on socio‐
hydrology: Socio‐hydrologic modeling: Tradeoffs, hypothesis testing, and validation. Water
Resources Research, 51(6), 4806-4814.
van Tol, J., Bieger, K., & Arnold, J. G. (2021). A hydropedological approach to simulate
streamflow and soil water contents with SWAT+. Hydrological Processes, 35(6), e14242.
Vogel, R. M., Lall, U., Cai, X., Rajagopalan, B., Weiskel, P. K., Hooper, R. P., & Matalas, N. C.
(2015). Hydrology: The interdisciplinary science of water. Water Resources Research, 51(6),
4409-4430.
Wagner, P. D., Bieger, K., Arnold, J. G., & Fohrer, N. (2022). Representation of hydrological
processes in a rural lowland catchment in Northern Germany using SWAT and SWAT+.
Hydrological Processes, 36(5), e14589.
Walters, C. (1997). Challenges in adaptive management of riparian and coastal ecosystems.
Conservation ecology, 1(2).
Chapter 3a
122
Xu, L., Gober, P., Wheater, H. S., & Kajikawa, Y. (2018). Reframing socio-hydrological research
to include a social science perspective. Journal of hydrology, 563, 76-83.
Xu, L., Gober, P., Wheater, H. S., & Kajikawa, Y. (2018). Reframing socio-hydrological research
to include a social science perspective. Journal of hydrology, 563, 76-83.
Zessner, M., Schönhart, M., Parajka, J., Trautvetter, H., Mitter, H., Kirchner, M., ... & Schmid, E.
(2017). A novel integrated modelling framework to assess the impacts of climate and socio-
economic drivers on land use and water quality. Science of The Total Environment, 579, 1137-
1151.
Zhao, Q., Chen, Y., Gone, K. P., Wells, E., Margeson, K., & Sherren, K. (2023). Modelling cultural
ecosystem services in agricultural dykelands and tidal wetlands to inform coastal infrastructure
decisions: A social media data approach. Marine Policy, 150, 105533.
Chapter 3b: A Socio-Hydrological Framework for Incorporating Farmers’
Behavioral Model into the SWAT+ Model and Implications for Improving Ecosystem
Health
Mahesh R Tapasa*, Randall Etheridgeb, Gregory Howardc, Matthew Maird
a Integrated Coastal Program, East Carolina University, Greenville, NC 27858, USA
bDepartment of Engineering, Center for Sustainable Energy and Environmental Engineering, East
Carolina University, Greenville, NC 27858, USA
c Department of Economics, East Carolina University, Greenville, NC 27858, USA
d Department of Economics, Appalachian State University, Greenville, NC 28608
*Corresponding author: tapasm21@students.ecu.edu (Mahesh R Tapas).
ORCID: 0000-0001-8833-5531
Abstract:
Eutrophication poses a significant threat to ecosystems, necessitating effective policies to mitigate
nutrient excess. This study adopts an interdisciplinary approach, merging concepts from
econometric and engineering frameworks to develop a socio-hydrological model. Integrating a
farmers' behavioral model into the hydrological model, the study surveyed around 279 farmers in
the Tar-Pamlico River Basin to gauge their responses to policy changes of cover crop adoption and
Chapter 3b
124
reduced fertilizer use. Analysis revealed that the majority of farmers are hesitant to adopt these
practices based on farmers' Willingness To Accept (WTA). Further, simulations using the Soil and
Water Assessment Tool Plus (SWAT+) software showed no significant reduction in nitrate loading
to the Pamlico Estuary from the widespread adoption of these practices, with both cover crops and
reduced fertilizer application failing to produce noticeable changes, underscoring the limitations
of the modeled Best Management Practices (BMPs) at the watershed scale.
The study highlights the complexity of nutrient transport in coastal watersheds and suggests
exploring combined BMPs and integrating socio-hydrological modeling for more accurate
predictions of agricultural policy impacts. It emphasizes the need to consider alternative nitrate
sources and scale-up interventions to effectively address eutrophication. By bridging econometric
and engineering frameworks, this research provides insights into the challenges of implementing
sustainable agricultural practices and underscores the importance of interdisciplinary approaches
in tackling environmental issues.
Study Area:
Tar-Pamlico Watershed, North Carolina, United States.
Keywords:
Socio-hydrological modeling; Best Management Practices (BMPs); Farmers' behavioral model;
SWAT+; Eutrophication; Ecosystem health; Nitrate load.
Chapter 3b
125
Highlights:
1. Integrating hydrological and economic models reveals complex dynamics in nutrient
management, emphasizing the need for interdisciplinary policy frameworks.
2. Despite incentives, farmers show limited willingness to adopt strict nitrogen limits,
highlighting economic and behavioral barriers in sustainable agriculture.
3. Hydrological models overestimate nitrogen reduction benefits by 15 times, underscoring
the need for models that incorporate socio-economic factors for accuracy.
Graphical Abstract:
Chapter 3b
126
1. Introduction:
In recent decades, anthropogenic climate change — climate change induced by human activities
— has substantially altered watershed dynamics worldwide (Chang et al., 2015; Palmer et al.,
2008). This change is due to human activities such as burning fossil fuels (Höök et al., 2013),
excessive use of fertilizers (Celikkol et al., 2017), development of water diversion networks (Bai
et al., 2023), and intensive agriculture (Lal 2013). Rising temperatures (Luo et al., 2013), shifting
precipitation patterns (Coopersmith et al., 2014), and the increasing frequency and intensity of
extreme weather events (Hettiarachchi et al., 2018), such as hurricanes (Ouyang et al., 2022) and
droughts (Peterson et al., 2021), are influencing watershed hydrology by altering river flow
regimes (Gibson et al., 2005), groundwater recharge rates (Nolan et al., 2007), and soil moisture
content (Hawley et al., 1983). These hydrological changes can intensify the issues of excessive
nutrients in water bodies (Petry et al., 2002), primarily from agricultural runoff (Cherry et al.,
2008) and urban development (Reckhow et al., 1999). Higher temperatures can enhance the
biological demand for oxygen in water bodies, exacerbating the effects of nutrient pollution, such
as accelerated eutrophication (Cross & Summerfelt 1986). This process can lead to harmful algal
blooms (Wurtsbaugh et al., 2019), hypoxia (Howarth et al., 2011), and loss of biodiversity (Hautier
et al., 2009), undermining the health of aquatic ecosystems and affecting water quality for
recreation and wildlife (Kapsalis & Kalavrouziotis, 2021).
Implementing agricultural Best Management Practices (BMPs), such as the strategic use of
cover crops and the reduction of chemical nitrogen (N) fertilization rates, offers a targeted
approach to mitigating the threat of excess nutrient loads in watersheds (Howard et al., 2024).
Cover crops, planted during times when the soil might otherwise be bare, play a crucial role in
improving soil structure, enhancing water infiltration, and increasing soil organic matter (Dabney
Chapter 3b
127
et al., 2001). This leads to reduced runoff and erosion, thereby limiting the flow of nutrients into
adjacent water bodies. Moreover, certain cover crop species have the ability to capture residual
nitrogen from previous crop fertilization. This can effectively reduce nitrate leaching into
groundwater and surface waters (Wendling et al., 2016). Reducing the application rate of nitrogen
fertilizers directly addresses a source of nutrient pollution (Shibu et al., 2010). This practice not
only reduces the risk of nitrate runoff into aquatic ecosystems but also encourages more efficient
use of fertilizers, resulting in economic benefits for farmers (Howard et al., 2024). Together, these
BMPs have the potential to contribute significantly to reducing excessive nutrients in agriculturally
dominated watersheds, thus improving water quality and supporting the health of aquatic
ecosystems as we adapt to the impacts of climate change (Dabney et al., 2001; Shibu et al., 2010).
Hydrological modeling, particularly using advanced tools like Soil and Water Assessment
Tool Plus (SWAT+), plays a pivotal role in evaluating the efficiencies of agricultural BMPs. By
simulating complex watershed processes, SWAT+ can assess how different BMP scenarios impact
water quantity and quality across various spatial and temporal scales (Singh et al., 2023). This
includes the ability to model the effectiveness of cover crops in reducing runoff and enhancing
infiltration, as well as quantifying the impact of reduced nitrogen application rates on nitrate
leaching into water bodies. SWAT+ allows for the exploration of BMP placement and timing under
varying climatic and land use conditions, offering insights into how these practices can be best
implemented to achieve maximum environmental benefits (Naganur et al., 2024). The model's
capacity to predict environmental outcomes of specific agricultural practices provides a valuable
decision-support tool for policymakers, land managers, and farmers aiming to mitigate nutrient
pollution and improve watershed health in a changing climate (Naganur et al., 2024; Singh et al.,
2023).
Chapter 3b
128
Incorporating stakeholders' perspectives into hydrological modeling is essential to ensure that
model outputs align with real-world conditions and stakeholder needs (Veisi et al., 2022). While
tools like SWAT+ have been proven to simulate physical processes, they often fall short of fully
capturing the socioeconomic factors that influence the adoption of BMPs by land managers
(Howard et al., 2024). Through the integration of stakeholder engagement and econometric
modeling, particularly farmer behavioral models, model simulations can be grounded in practical
realities and enhance the accuracy of model predictions. This methodology facilitates the
identification of socially acceptable and economically viable BMPs, leveraging statistical methods
to quantify how economic, environmental, and policy factors affect farmers' decision-making.
Ultimately, this has the potential to spur wider adoption of sustainable practices and more effective
watershed management strategies, bridging the gap between physical process simulation and the
socio-economic dynamics of agricultural practice adoption (Behboudian et al., 2023; De Bruijn et
al., 2023).
In response to the outlined challenges, this study will undertake a multifaceted approach that
combines nitrate modeling with an evaluation of the effectiveness of cover crops and reduced
nitrogen application rates in mitigating downstream nutrient loading. Recognizing the importance
of socio-economic factors in the adoption and success of BMPs, this research will also integrate
stakeholder views through farmers' behavioral models. This will ensure that the modeling not only
reflects physical and biological processes but also incorporates the decision-making behaviors of
land managers. Finally, by comparing the outcomes of traditional hydrological modeling with
those informed by a farmer behavioral model, this study aims to highlight the added value of
embedding social dynamics into hydrological models. Such an integrative approach should
enhance the accuracy of model predictions and provide actionable insights for policymakers,
Chapter 3b
129
ultimately contributing to more effective and sustainable watershed management strategies in the
face of ongoing climate change.
1.1. Study objectives:
The objectives of this study are: 1) To develop and optimize a hydrological model for the Tar-
Pamlico watershed in coastal North Carolina using SWAT+; 2) To implement cover crops and
reduced N-fertilizer application rate and find their impact on nitrate loading at Pamlico estuary
using the hydrological model; 3) To survey farmers in the watershed to identify their responses to
policy changes; 4) To develop a farmers' behavioral model to identify their willingness to accept
cover crops and reduced N-fertilizer application under the current EQIP rate; 5) To integrate the
hydrological model (engineering framework) and the farmers' behavioral model (econometric
framework) to develop a socio-hydrological model; 6) To test the impacts of different policies on
ecosystem health using this socio-hydrological model.
2. Materials & Methodology:
2.1 Study Area
In this study, we applied our socio-hydrological modeling framework to the Tar-Pamlico
watershed (Figure 3b.1), a coastal area in eastern North Carolina covering 16,576 km2. This
watershed, characterized by its diverse land use including agriculture (27.9%), forests (33.9%),
and wetlands (31.9%), supports a variety of crops with soybeans (40%), corn (19%), and cotton
(19%) the predominate agricultural crops. With a population of 470,000, the Tar-Pamlico
watershed plays a significant role in discharges to the Pamlico estuary, an area currently facing
challenges with algae blooms attributed to excessive nitrate levels. By implementing our
Chapter 3b
130
methodology in this watershed, we aim to explore the ability of different BMPs to address the
critical issue of elevated nutrient loads, thereby contributing to the overall health of the ecosystem
and mitigating the adverse effects on the Pamlico estuary.
Figure 3b.1 Study Area (Tar-Pamlico watershed)
2.2 Hydrological model: Engineering Framework
We used the SWAT+ tool for hydrological modeling. SWAT+ is designed to provide
comprehensive modeling of watershed and sub-watershed dynamics (SWAT+ IO Document,
2020). It serves as a critical tool for decision-making in water resource management, agricultural
planning, and environmental conservation, offering insights into the effects of land management
Chapter 3b
131
practices on water quality and agricultural productivity (Bieger et al., 2017; Wu et al., 2023). By
enabling the simulation of complex environmental processes, SWAT+ supports policy
development and planning. SWAT+ uses a semi-distributed hydrological framework and enhanced
spatial flexibility compared to SWAT (SWAT+ IO Document, 2020). Particularly noteworthy is
its capability to simulate flow and nutrient dynamics, including detailed analyses of plant yield,
denitrification rates, and nitrate loss in groundwater, surface water, and lateral flows. This
precision allows for the evaluation of the impacts of climate change and land use changes on water
resources, making SWAT+ a great tool for future resource planning and environmental protection
efforts (SWAT+ IO Document, 2020).
For SWAT+ model setup, we utilized various data sources to represent watershed
characteristics accurately. Watershed boundary shapefiles were obtained from the United States
Geological Survey (USGS) StreamStats website. The model was initialized using Digital Elevation
Model (DEM) data (year: 2011), land-use and land-cover data from the National Land Cover
Database (NLCD) (year: 2008), and soil data from the Soil Survey Geographic Database
(SSURGO) (year: 2015), all at a 90m resolution. These datasets were integrated into QSWAT+ to
delineate streams, sub-basins, and Hydrological Response Units (HRUs), essential for simulating
the complex hydrological processes within the watershed. We included atmospheric deposition
data from the National Atmospheric Deposition Program (NADP, 2023). We integrated crop data
and fertilizer application rates into the SWAT+ model setup based on Tar-Pamlico watershed
reports from the North Carolina Department of Environmental Quality (NCDEQ) and the North
Carolina Department of Agriculture & Consumer Services (NCDEQ, 2014; NCGAR, 2022). These
datasets provided detailed information on agricultural practices within the watersheds, including
types of crops grown, the extent of winter cover crop adoption, and specific fertilizer usage rates.
Chapter 3b
132
Incorporating this information was crucial for more accurately simulating agricultural runoff and
nutrient loading, allowing for a more comprehensive assessment of the watershed's hydrology and
nutrient dynamics. The Tar-Pamlico watershed contains a total of 22 Wastewater Treatment Plants
that were included in the model. We retrieved flow and nitrate data for each Wastewater Treatment
Plant from National Pollutant Discharge Elimination System (NPDES) (National Pollutant
Discharge Elimination System (NPDES) | US EPA). All these details were added to increase the
accuracy of flow and nitrate simulations.
We obtained observed flow data from USGS stations at Greenville (02084000) and at
Washington (02084472). Flow data at Washington station is available only before September 2006.
As our nitrate data is available in Washington, we extrapolated our flow data at the Washington
station based on flow at the Greenville station because there was a strong linear correlation. The
linear regression equation (Flow at Washington = 1.0176 × flow at Greenville + 27.5774) with an
R2 of 0.90 indicate that approximately 90% of Washington's discharge variability can be estimated
from Greenville's data, offering a reliable method for compensating the data scarcity at
Washington. Nitrate concentration data were obtained from the North Carolina Department of
Environmental Quality (NCDEQ). We calculated monthly nitrate load at Washington using nitrate
concentrations and extrapolated flows (Appendix 1). Because SWAT+ cannot handle the
bidirectional flow (SWAT+ IO Document, 2020), we converted any negative flows (0.39%) from
our observed flow data file to zero. This ensured that we simulated flow as low as possible during
the days of backflow.
2.2.1. SWAT+ model calibration
We used the SWATrunR package (Schuerz 2019) to calibrate the SWAT+ model, with a focus on
simulating flow and nitrate load at the Washington station. Our calibration efforts focused on a set
Chapter 3b
133
of 21 parameters crucial for accurately representing hydrological processes based on sensitivity
analysis (Appendix 1) and the existing literature. These parameters were distributed across the
basin, soil, HRU, and aquifer levels, with 6 basin parameters, 4 soil parameters, 10 HRU
parameters, 2 plant parameters, and 1 aquifer level parameter as shown in Table 3b.1. We
customized an R script, using the SWATrunR package as a foundation (Appendix 1).
We employed the parameter adjustment methods of absolute change (x' = x + y), percent
change (x' = y * x / 100), and absolute value (x' = y), where x represents the default value, x'
denotes the new value, and y signifies the calibrated parameter value. The selection of these
adjustment types depended on both the parameter's resolution and its initial range. For instance,
when dealing with basin-level parameters, we predominantly utilized absolute value adjustments.
Additionally, the choice between absolute and percentage changes was determined by the
magnitude of the parameter's range: for narrow ranges, absolute changes were preferred, while
percentage changes were favored for wider ranges. This approach facilitated targeting a broad
range of parameter combinations during model calibration, ensuring flexibility to adjust
parameters effectively across varying resolutions and initial ranges (Table 3b.1).
Chapter 3b
134
Table 3b.1 SWAT+ model optimization parameters details [bsn: Basin, sol: Soil, hru: Hydrological Response Unit, plt: Plant]
Parameter
Parameter
Range (Unit)
Resolution
Type of
Change
Description
Initial range
used for model
calibration
(min, max)
Calibrated
parameter
value
surlag
0.05, 24.0
(days)
bsn
Absolute
Value
Surlag Controls delay in surface runoff release
0.05, 24
3.574
cmn
0.001, 0.003
(-)
bsn
Absolute
Value
Rate factor for humus mineralization of organic nutrients
0.001, 0.003
0.0018
cdn
0.0, 3.0
(-)
bsn
Absolute
Value
Denitrification rate control
0, 3
2.261
sdnco
0.0, 1.0
(-)
bsn
Absolute
Value
Denitrification threshold water content
0, 1
0.559
nperco
0.0, 1.0
(-)
bsn
Absolute
Value
Nitrate percolation coefficient
0.01, 1
0.080
n_updis
0.0, 100.0
(-)
bsn
Absolute
Value
Nitrogen uptake distribution parameter controlling depth
distribution of nitrogen uptake in soil
0, 100
49.233
awc
0.01, 1.0
(mm H2O mm-1)
sol
Absolute
Change
The difference in soil water content between field capacity
and permanent wilting point
-0.3, 0.3
0.081
bd
0.9, 2.5
(g cm-3)
sol
Absolute
Change
Moist bulk density, representing soil's mass-to-volume ratio
at or near field capacity.
-0.4, 0.8
0.271
k
0.0001, 2000.0
(mm hr-1)
sol
Percent
Change
Saturated hydraulic conductivity, indicating the ease of
water movement through soil
-30, 30
-4.024
z
0.0, 3500.0
(mm)
sol
Percent
Change
Depth from soil surface to bottom of layer
-30, 30
1.469
esco
0.0, 1.0
(-)
hru
Absolute
Change
Soil evaporation compensation factor which allows
modification of depth distribution to meet soil evaporative
demand, considering capillary action, crusting, and cracks.
-0.3, 0.3
-0.069
epco
0.0, 1.0
(-)
hru
Absolute
Change
Plant uptake compensation factor which allows adjustment
of water uptake depth distribution in response to plant
transpiration demand and soil water availability.
-0.3, 0.3
-0.114
biomix
0.0, 1.0
(-)
hru
Absolute
Change
Biological mixing efficiency, determining redistribution of
soil constituents by biota activity.
-0.3, 0.3
-0.048
latq_co
0.0, 1.0
(-)
hru
Absolute
Change
Coefficient for the Plant ET curve number
-0.3, 0.3
0.101
Chapter 3b
135
perco
0.0, 1.0
(fraction)
hru
Absolute
Change
Percolation coefficient, adjusting soil moisture for
percolation to occur.
-0.3, 0.3
-0.272
cn2
35.0, 95.0
(-)
hru
Percent
Change
Curve number for Condition II runoff potential.
-30, 30
12.283
cn3_swf
0.0, 1.0
(-)
hru
Percent
Change
Soil water factor for the curve number for condition III
runoff potential
-30, 30
-24.729
ovn
0.01, 30.0
(-)
hru
Percent
Change
Manning's "n" value for overland flow velocity estimation
-30, 30
14.865
canmx
0.0, 100.0
(mm H2O)
hru
Percent
Change
Maximum canopy storage, representing the maximum
amount of water held in the canopy when fully developed.
-30, 30
-0.189
lat_ttime
0.5, 180.0
(days)
hru
Percent
Change
Lateral flow travel time allows the model to calculate travel
time based on soil hydraulic properties.
-30, 30
-19.100
revap_min
0.0, 50.0
(m)
aqu
Percent
Change
Threshold depth of water in shallow aquifer for percolation
to deep aquifer
-30, 30
7.659
Lai_pot
0.5, 10
(m2 m-2)
plt
Absolute
Value
Potential maximum leaf area index
NA
Corn: 5
Cotton: 2.5
Soyb: 2.027
Harv_idx
0.01, 1.25
(-)
plt
Absolute
Value
Harvest index- crop yield/aboveground biomass
NA
Soyb: 0.418
Chapter 3b
136
2.2.2. Best Management Practices Simulation
In this study, we simulated two commonly used best management practices: cover crops and
reduced fertilizer application rates. We used winter wheat as the cover crop, which is commonly
used in the Tar-Pamlico Basin. We planted the cover crop 14 days after harvesting the main summer
cash crop. We terminated the cover crop prior to planting the summer crop the following spring.
We applied a 30% reduction in nitrogen fertilizer use to simulate policies related to strict nitrogen
application.
2.3. Farmers’ behavioral model: Econometrics framework
We conducted a survey (Appendix 1) among farmers in the Tar-Pamlico River basin and other
coastal areas in eastern North Carolina to gauge farmers’ interest in voluntary conservation
programs. Our goal was to use survey results to parameterize a behavioral model that predicts
farmers' responses to various policy changes, especially concerning the provision of payments for
cover crops and reductions in fertilizer usage. By incorporating scenarios that reflect aspects of
existing and hypothetical agricultural support programs (i.e. Environmental Quality Incentives
Program, or EQIP), the survey sought to capture farmers' perspectives on and likely response to
potential policies and economic incentives, thereby bridging the gap between policymakers and
farmers.
The survey asked farmers about their specific farming practices, environmental concerns, and
the potential impact of policy changes on these practices. This comprehensive approach was
intended to gather insights into farmers' current challenges and their willingness to accept different
practices in light of new policies or economic conditions. These surveys are increasingly common
in agri-environmental policy analysis (Howard & Roe, 2013; Ma et al., 2012; Peterson et al., 2015).
Chapter 3b
137
2.3.1. Choice Experiment Design in Farmer Preference Survey
The survey incorporated a choice experiment to gauge farmers' preferences for various
hypothetical voluntary conservation contracts, aimed at reducing nutrient runoff (Figure 3b.2).
This experiment varied five key attributes across different scenarios: cover crop usage (required
or not), strict nitrogen application (mandatory or optional), lenient nitrogen application (mandatory
or optional), funding agency (State and Federal government agencies or private conservation
groups), and the annual payment amount ($10, $40, $70, $100, and $130 per acre; or $25, $99,
$173, $247, and $321 per Ha). Each respondent was presented with two scenarios, each containing
two conservation contracts and a "neither" option, allowing us to simulate real policy decisions
(Figure 3b.2).
Chapter 3b
138
Figure 3b.2 Sample cost share contract
We sent surveys to over 1870 farmers through a combination of solicitation via email, solicitation
via postal mail, and direct engagement at farmer meetings in eastern North Carolina. Our primary
method of contact involved mailings to farmer addresses. We utilized a mixed-mode web push
approach that included an initial invitation to participate as well as a follow-up mailing to initial
Scenario 1
Please consider the terms of Programs A & B below for your field and answer the questions
that follow as if a real conservation contract was being offered to you.
Program A
Program B
Limited Nitrogen Application
Apply No more than 40
lbs. per acre
Apply No more than 60
lbs. per acre
Cover Crops (Planting a legume crop
after harvesting the main cash
crop)
Must be planted after
2023 harvest
No Requirement
Funding Source (who is providing the
money and enforcing the
contract)
Private Conservation
Group
State and Federal
Government Agencies
Annual Cost Share Payment to You
$130/acre
$10/acre
28. Which program do you prefer?
1 = Program A 2 = Program B 3 = Neither program
29. How Certain are you about your choice?
1 2 3 4 5 6 7 8 9 10
Not at all Certain Very Certain
Chapter 3b
139
non-responders. This follow-up mailing included a paper survey and a pre-paid return envelope
(Dillman et al., 2014). We received 279 responses, with 76 of these providing enough data to
include them in the construction of the farmers' behavioral model. The vast majority of responses
that did not provide sufficient data involved respondents indicating that they were no longer
actively farming. These 76 responses, which represent up to 29 km2 of farmland, are particularly
valuable for developing a model that accurately reflects farmer willingness to accept different
agricultural practices under a voluntary conservation contract setting.
The provided STATA code (Appendix 1) outlines an econometric analysis of the behavioral
responses of farmers in North Carolina, grounded in the Random Utility Maximization (RUM)
theory (Habib 2011). RUM theory specifies that the utility farmer i receives from, choosing
alternative j in time period t, denoted Uijt, can be described as a function of the observed attributes
Xijt and unobserved factors ϵijt (Howard et al., 2024). Thus, the utility function can be expressed as
follows:
Uijt=Xijtβi+ϵijt
Where:
• Xijt represents the attributes observed, including state funding presence, payment rate, strict
nitrogen regulation, lenient nitrogen regulation, an alternative-specific constant for the “no
contract” option, and cover crops utilization.
• βi is a vector that represents preference parameters for the population, with the subscript i
indicating that parameters may vary across individuals, thus accounting for heterogeneity
in preferences among farmers.
Chapter 3b
140
• ϵijt is the unobserved error term.
To incorporate preference heterogeneity directly into the model, we utilize the mixed logit
model. In this model, individual-specific deviation parameters ηi may be added:
Uijt=Xijt(β+ηi)+ϵijt
This means that individual i's preference parameter, βi is equal to the sum of the population
average preference parameter a vector β and the individual-specific deviation parameter. For each
attribute with preference heterogeneity, the model estimates a set of two parameters, one for the
average preference in the population and one for the standard deviation of the individual deviation
parameters (Gaur et al., 2023).
Assuming the error term follows a Type-1 extreme value distribution with mean zero, the
probability Pi{c1,c2,...cT} of individual i's sequence of choices [c1,c2,...cT] over T time periods (Gaur
et al., 2023) is given by:
where:
• Ximt represents the observed attributes of the alternatives available to farmer i in choice
situation t, which includes factors from the survey such as the presence of state funding,
payment rates, strict and lenient nitrogen application regulations, and the requirement of
cover crops.
Chapter 3b
141
• βi is a vector of coefficients associated with the attributes Ximt, reflecting the mean
preferences of the farmers but allowing for individual-specific deviations to capture
preference heterogeneity.
• f(βθ) is the density function of β, with θ being the parameters of the distribution, indicating
how the preference parameters β are distributed in the population of farmers.
• M is the number of alternatives available in the choice set.
In refining our econometric model, we focused on allowing preference heterogeneity for the
alternative specific constant (ASC) and cover crop utilization. Attempts to include heterogeneity
for nitrogen application restrictions did not converge, suggesting limited variability in preferences
for these attributes among the farmers surveyed.
(3)
2.3.2. Estimation of Willingness to Accept (WTA)
In our survey of the farmers in the Tar-Pamlico basin, we presented choices between two
hypothetical agricultural assistance programs alongside a status quo option of no contract. This
enabled us to examine the impact of specific regulatory or funding conditions on farmers'
willingness to implement environmentally beneficial practices. Utilizing the mixed logit model
described in the previous section, we analyzed how these variables related to policy interventions,
economic incentives, and environmental practices influence farmer choices, thus revealing insights
into the decision-making processes that drive the adoption of sustainable agricultural practices
(Howard et al., 2024).
Chapter 3b
142
The estimation of WTA is a critical element of our econometric framework, serving to gauge
the minimum amount a farmer would accept to alter their current agricultural practice to
incorporate a practice that is potentially less profitable but more environmentally friendly. It is
derived from the utility differences provided by the mixed logit model's coefficients and is
expressed mathematically as the monetary compensation required to offset the utility loss incurred
by switching from the farmer’s current status-quo to a new situation where they accept an offered
conservation contract. The WTA for policy j is calculated as:
In this formula, Uij represents the utility from accepting the conservation contract, Ui baseline
is the utility for the current practice, and payment, βpayment,i is the coefficient for the policy payment
attribute, which also represents the marginal utility of income changes.
Our econometric analysis captures individual-specific preference deviations and the inherent
randomness in farmers' responses to incentives, thus highlighting unique decision-making
processes. The mixed logit model provides coefficients and standard errors for each contract
attribute, which are essential for determining the marginal effects on farmer utility and calculating
the likelihood of contract acceptance and overall Willingness to Accept (WTA). In essence, each
farmer is assigned a draw from the distribution of farmer preferences estimated in the farmer
behavioral model, so each farmer can potentially have a different simulated minimum WTA for a
specified contract. These insights are vital for developing targeted policies and interventions that
promote sustainability and economic efficiency, enhancing our understanding of agricultural
practices and supporting resilience in farming communities (Howard et al., 2024).
Chapter 3b
143
2.4. Hydrological-Econometric models integration
Integrating hydrological and econometric models is a crucial step toward achieving a
comprehensive understanding of watershed dynamics and agricultural decision-making processes
(Liu et al., 2008). By combining the insights gained from the farmers’ behavioral model with the
SWAT+ hydrological model, we aim to develop an integrated framework that bridges the gap
between policy interventions and on-the-ground agricultural practices (Du et al., 2020). By
analyzing farmers' willingness to adopt these practices (cover crops and reduced N-rate
application) based on the incentives offered (standard EQIP cost share rates), we aim to estimate
the extent of land conversion for cover crops and reduced fertilizer application. This integration
allows us to assess the effectiveness of incentive programs in promoting sustainable agricultural
practices and inform policymaking for environmental conservation (Cheng & Li, 2015).
Table 3b.2 outlines seven distinct scenarios simulated in this study. Scenario 1 serves as the
baseline, involving no implementation of additional Best Management Practices (BMPs). Utilizing
the findings from Scenario 1, we identified the top 12% of Hydrological Response Units (HRUs)
for highest agricultural surface nitrate export for Scenarios 2 and 4, as well as the top 24% of
HRUs for highest agricultural surface nitrate export for Scenarios 3 and 5. From an engineering
perspective, these HRUs are considered the most optimal areas for BMP deployment to minimize
the watershed's nitrate export effectively (Sheshukov et al., 2016).
Scenarios 2, 3, and 6 explore the impacts of cover crops, while Scenarios 4, 5, and 7 examine
the effects of a 30% reduction in N-fertilizer application rates across varying percentages of the
Tar-Pamlico watershed. Scenarios 2-5 provide conservation contracts at current EQIP rates ($50
per acre for N reduction, $75 per acre for cover crops) and simulate perfect adoption and land use
change in the targeted HRUs. Scenarios 6 and 7 similarly offer the same conservation contracts to
Chapter 3b
144
the same targeted HRUs but use a farmer behavioral model to predict the proportion of farmers
who will accept the offer and consequently, what percentage of HRUs will be converted. A
comprehensive cost-benefit analysis was conducted to assess the expenditure associated with
implementing these practices and the benefits gained through nitrogen loss reduction.
Table 3b.2 Policy scenarios for hydrological and socio-hydrological modeling
Policy Scenario
Type of model
Description
Scenario 1
Hydrological Model
Baseline
Scenario 2
Hydrological Model
Cover crops applied on 12% of the total
agricultural land
Scenario 3
Hydrological Model
Cover crops applied on 24% of the total
agricultural land
Scenario 4
Hydrological Model
30% reduced N fertilizer application on 12% of
the agricultural area
Scenario 5
Hydrological Model
30% reduced N fertilizer application on 24% of
the agricultural area
Scenario 6
Socio-Hydrological
Model
Cover crops applied to areas based on the
farmers' behavior model.
Scenario 7
Socio-Hydrological
Model
30% reduced N fertilizer application to areas
based on the farmers' behavior model.
Chapter 3b
145
3. Results and Discussion
3.1. SWAT+ model optimization
In this study, we evaluated the performance of the model for simulating monthly nitrate load at the
Washington, NC station over two distinct periods (Figure 3b.3), a calibration period from 2003 to
2011 and a validation period from 2012 to 2019. The model's ability to simulate nitrate loads
similar to the observed values was assessed using three key performance indices: Nash-Sutcliffe
Efficiency (NSE), the Coefficient of Determination (R²), and the Kling-Gupta Efficiency (KGE).
During the calibration period, the model demonstrated a good level of accuracy in simulating
nitrate loads, considering Tar-Pamlico is a coastal watershed (Upadhyay et al., 2022), achieving
an NSE of 0.61, an R² of 0.61, and a KGE of 0.77. These metrics indicate a strong agreement
between the observed and simulated data, suggesting that the model is well-calibrated and capable
of capturing the key dynamics governing nitrate transport and retention within the Tar-Pamlico
watershed (Upadhyay et al., 2022).
Figure 3b.3 Simulated and observed nitrate load for calibration and validation
Chapter 3b
146
For the validation period, the performance indices indicated a moderate level of model
accuracy, with an NSE of 0.33, an R² of 0.33, and a KGE of 0.39. While these values are lower
than those observed during the calibration period, they still demonstrate the model's capacity to
provide reasonable nitrate load predictions under varying conditions. The decline in performance
indices during the validation period was likely due to uncertainties associated with the reduced
frequency of observed data from 2012 to 2019 (Birgand et al., 2011). After 2010, the frequency of
nitrate concentration sampling by NCDEQ shifted from daily to monthly. Between January 2003
and December 2011, a total of 1,134 nitrate concentration measurements were collected, compared
to only 88 data points during the validation period from January 2012 to December 2019.
3.2. Farmers Behavioral Modeling:
The results from the mixed logit model (Table 3b.3), based on 76 farmer survey responses,
demonstrate statistically significant factors that influence farmers' choices regarding conservation
contracts. Notably, the payment variable (Coef. = 0.0321, P < 0.000) indicates that money is a
highly significant factor in the farmer’s decision to choose a conservation contract. Higher
payment rates increase farmer utility and the probability of farmers accepting these contracts. This
finding underscores the importance of financial incentives in promoting widespread conservation
practices among farmers. The status-quo, or no contract, alternative specific constant (ASC) is also
significant (Coef. = 3.161, P = 0.003), suggesting a general bias among farmers against choosing
a conservation contract.
Chapter 3b
147
Table 3b.3: Mixed Logit Model Results
Variable
Coefficient
Standard
Error
z-
value
P-
value
95% Confidence
Interval
Mean
State funding presence
.0108
.535
0.02
0.984
-1.037, 1.059
Payment rate ($)
.032
.009
3.54
0.000
0.014, 0.050
Strict nitrogen regulation
(lb/ acre)
-2.134
.869
-2.45
0.014
-3.838, -0.430
Lenient nitrogen
regulation
(lb/ acre)
-1.11
.873
-1.27
0.203
-2.824, 0.599
Alternative specific
constant
3.161
1.046
3.02
0.003
1.111, 5.211
Cover crops utilization
-.845
.971
-0.87
0.384
-2.749, 1.059
Standard Deviation
Alternative specific
constant
3.35731
1.188573
2.82
0.005
1.028, 5.687
Cover crops utilization
2.849883
1.294219
2.20
0.028
0.313, 5.387
Chapter 3b
148
Conversely, the coefficient for the dummy variable representing strict nitrogen application
limits (StrictNitrogenRegulation, Coef. = -2.134354, P = 0.014) is negative and statistically
significant, indicating that contracts with stringent nitrogen application restrictions are less likely
to be chosen by farmers. This result may highlight concerns among farmers about the feasibility
or economic implications of significantly reducing nitrogen use on their crops. The standard
deviations for ASC (SD = 3.357, P = 0.005) and cover crops (SD = 2.8499, P = 0.028) are
significant, illustrating variability in farmers' preferences for these contract attributes. This
heterogeneity suggests that while some farmers may be more inclined towards contracts that
include these features, others may have a strong preference against them, highlighting the need for
flexible policy designs to accommodate diverse farmer preferences (Howard et al., 2024).
Although some of the management practice attributes do not achieve statistical significance, all
mean coefficient estimates maintain the expected negative sign. This indicates that, all else equal,
farmers are reluctant to implement these practices. Notably, the negative coefficient for strict
nitrogen (N) limits is larger than that for lenient N limits, aligning with expectations that stricter
regulations are less appealing (Howard et al., 2024). Additionally, the high p-values associated
with some coefficients may be attributable to the small sample sizes involved in the study. Indeed,
we see that, while cover crops are not the norm across fields, 11.7% of our sample report using
cover crops currently, and in other watersheds cover crops are voluntarily adopted by a minority
of farmers (Burnett et al., 2015).
In contrast, the variable representing state funding (dummystatefunding, Coef. = .0107, P =
0.984) is not statistically significant, suggesting that this factor does not have a discernible impact
on farmers' contract choices within the context of this model. Overall, these results indicate that
financial incentives and the specifics of conservation practices (such as nitrogen application limits)
Chapter 3b
149
play crucial roles in influencing farmers' decisions to participate in conservation programs. The
heterogeneity in preferences highlighted by the significant standard deviations underscores the
importance of designing flexible conservation programs that can cater to a wide range of farmer
needs and concerns (Howard et al., 2024).
In the simulation for strict nitrogen (N) restriction policies under the current EQIP rate of $50
per acre or $123.5 per Ha, our analysis reveals a modest enrollment projection. Based on the mixed
logit model's outcome and subsequent simulations, it is estimated that only 3.75% of the farmers
in our sample would opt to participate in a policy that mandates strict nitrogen application limits
under current EQIP rates. The process involved calculating the WTA strict N restrictions at or
below a $123.5 per Ha rate, then aggregating the total area farmers are willing to enroll under such
conditions. The low percentage of potential enrollees underscores a cautious or reluctant stance
among farmers toward adopting strict nitrogen application limits, even when incentivized.
The analysis regarding the adoption of cover crops under the current Environmental Quality
Incentives Program (EQIP) rates reveals a more optimistic participation forecast compared to strict
nitrogen restrictions. Our simulation indicates that approximately 26.2% of farmers would be
willing to enroll in a policy promoting cover crop practices when the WTA threshold is set at or
below $75 per acre, or $185.3 per Ha. However, our model suggests that not all of those who enroll
in the conservation program would count as new or additional acreage in the practice. While we
see a contract acceptance rate of 26.23%, some of these farmers report already utilizing cover crops
in the past, so our prediction for the increase in new acreage of the management practice is a more
modest 14.5%. This higher inclination towards cover crop adoption suggests that, unlike strict
nitrogen application restrictions, farmers perceive cover cropping as a more viable and potentially
beneficial conservation practice under the EQIP framework. The significant difference in
Chapter 3b
150
enrollment projections between these two conservation practices underscores the need for tailored
incentive structures that reflect the varying levels of farmer receptivity and the distinct
environmental benefits each practice offers (Rossi & Hinrichs, 2011).
Understanding the preferences, constraints, and incentives that influence farmers' decisions to
adopt conservation practices is vital for designing effective environmental policies (Liu et al.,
2018). These behavioral insights serve as foundational components for simulating policies using
hydrological models, ensuring that such models reflect realistic agricultural management scenarios
(Kuil et al., 2018).
3.3. Effects of reduced N-fertilizer application rate and cover crops on watershed processes
and N-load.
3.3.1. Effects of 30% reduced fertilizer application on nitrate load
The effect of a 30 percent reduction in fertilizer application across varying percentages of
agricultural land on the total nitrate load at Pamlico Estuary and Washington, NC, for engineering
models scenarios is shown in Figure 3b.4.
Chapter 3b
151
Figure 3b.4 Effects of reduced fertilizer on nitrate load for the baseline, 12% (scenario 4), and
24% (scenario 5) of total agricultural area in the watershed
At the Pamlico Estuary, the winter season exhibits the highest nitrate loads across the baseline,
scenario 4, and scenario 5 (Table 3b.4). In the spring, nitrate loads remain elevated, suggesting
significant seasonal runoff (Table 3b.4) and greater nitrate availability from agricultural activity
(Bruland et al., 2003). The fall and summer seasons show comparatively lower nitrate loads. As
the area with reduced nitrogen fertilization expands, there is a noticeable decrease in nitrate
loading, most prominently in the spring and winter seasons. This underscores the potential
effectiveness of nitrogen management strategies during periods of increased runoff (Table 3b.4).
In Washington, NC, similar patterns are observed, although with overall lower nitrate loads
than the Pamlico Estuary (Table 3b.4). The spring season shows the highest loads, followed by
winter, with noticeable reductions in nitrate loading as the area of nitrogen reduction increases.
Chapter 3b
152
The least variation across scenarios is seen in summer, suggesting a more stable nitrate input during
this season. The more stable nitrate input during summer in Washington, NC, is likely due to
consistent agricultural practices, less intense rainfall, increased biological uptake by plants, and
more uniform water flows, leading to lower variability in nitrate levels (Dukes & Evans, 2006)
across multiple scenarios (Table 3b.4).
Table 3b.4 Seasonal average flow and nitrate load (Kg N) with reduced fertilizer application rate
Location
Season
Baseline
12% Area N
reduction
24% Area N
Reduction
Flow
(m³s-1)
Pamlico
Estuary
Fall
221,900
219,700
218,100
116
Spring
527,700
515,200
506,100
124
Summer
274,400
266,700
264,100
89
Winter
736,300
734,600
733,100
155
Washington,
NC
Fall
138,300
136,800
135,600
57
Spring
331,100
325,100
315,100
87
Summer
158,900
155,700
154,200
44
Winter
290,600
289,100
287,800
94
Chapter 3b
153
To determine if the nitrate loading under all three scenarios (Baseline, scenario 4, and scenario 5)
at the Pamlico Estuary are statistically different from each other, we performed an Analysis of
Variance (ANOVA) on monthly nitrate load at Pamlico estuary (Table 3b.5). ANOVA is a statistical
method used to compare the means of three or more samples to see if at least one of them differs
significantly from the others. If the p-value from the ANOVA test is less than a significance level
(0.05), we can reject the null hypothesis that all groups have the same mean, suggesting that at
least one scenario differs in terms of nitrate loading (St & Wold, 1989).
Table 3b.5 Reduced fertilizer application ANOVA (single factor) test at Pamlico estuary
SUMMARY
Groups
Count
Sum
Average
Variance
Baseline
204
29923200
146682.3529
13625690057
Scenario-4
204
29514920
144680.9804
13416121673
Scenario-5
204
29263940
143450.6863
13284143139
ANOVA
Source of Variation
SS
df
MS
F
P-value
F crit
Between Groups
1085469390
2
542734694.8
0.040376083
0.960431
3.01051703
Within Groups
8.18617E+12
609
13441984956
Total
8.18725E+12
611
The ANOVA test yielded a p-value of 0.96, which is much greater than the significance level
of 0.05. This result indicates that we cannot reject the null hypothesis, suggesting there is not
Chapter 3b
154
enough statistical evidence (St & Wold, 1989) to say that the mean monthly nitrate loading under
the four scenarios at the Pamlico Estuary are significantly different from each other.
It's important to note that while the ANOVA test provides insight into the overall differences
across groups, it doesn't specify which groups differ from each other (St & Wold, 1989). However,
in this case, the high p-value suggests that the differences in mean nitrate loading between the
scenarios are likely not to be significant.
Despite implementing nitrogen reduction strategies across 12%, and 24% areas, the analysis
suggests that these measures, within the scope of the study, do not significantly alter the mean
monthly nitrate loading when compared to the baseline scenario with no nitrogen reduction (Table
3b.5). The lack of statistically significant differences in nitrate loading under various nitrogen
reduction scenarios at the Pamlico Estuary could be attributed to several factors (Barker et al.,
2008).
This outcome may indicate that the predominant sources of nitrate loading in the Pamlico
Estuary are not primarily agricultural runoff, or that the scale of reduction explored in the scenarios
is insufficient to manifest a statistically significant impact given the natural variability of nitrate
loading and other contributing factors. Additionally, excessive nitrate leaching from groundwater,
possibly from legacy nitrate, could be contributing significantly to the observed levels. The
effectiveness of nitrogen reduction strategies might also be influenced by the spatial distribution
of runoff sources and the timing of nitrogen applications in agricultural practices (Craswell, 2021).
3.3.2. Effects of Cover Crops on Nitrate Load
The results from multiple scenarios reflecting varying degrees of cover crop implementation,
aimed at reducing nitrate export to the estuary are shown in Figure 3b.5.
Chapter 3b
155
Figure 3b.5 Effects of reduced fertilizer on monthly nitrate load to the Pamlico estuary for the
baseline, 12% (scenario 2), and 24% (scenario 3) coverage with cover crops.
The seasonal average nitrate loads (kg N) under varying cover crop scenarios were assessed
at two locations: Pamlico Estuary and Washington, NC (Table 3b.6). In the Pamlico Estuary, winter
consistently showed the highest nitrate loads (734,200 to 734,700 kg N) in each scenario,
suggesting higher discharge (120 m3/s). Contrary to expectations, the decrease in nitrate loading
with the expansion of cover crop areas is not universally observed. For instance, during the spring,
an increase in nitrate loads (537,600 kg N) was noted with the highest cover crop coverage
(scenario 3, 24% area application of cover crops) compared to the baseline (527,700 kg N),
challenging the presumption of uniform effectiveness of cover crops in nitrate load reduction
across seasons. For Washington, NC, while echoing a similar trend of high nitrate loads in the
spring (327,600 to 331,100 kg N) and winter (288,100 to 290,600 kg N), also underscores a more
nuanced response to increased cover crop adoption. The changes in nitrate loading with varying
Chapter 3b
156
cover crop scenarios were modest, suggesting that the cover crops' role in mitigating nitrate runoff
is complex and influenced by many factors (Craswell, 2021).
Table 3b.6 Seasonal variation of flow and nitrate load (kg N) with cover crops
Location
Season
Baseline
12% Cover
Crops
24% Cover
Crops
Flow
(m³ s-1)
Pamlico
Estuary
Fall
221,900
220,600
221,400
116
Spring
527,700
537,200
537,600
124
Summer
274,400
281,000
281,900
89
Winter
736,300
734,200
736,700
155
Washington,
NC
Fall
138,300
138,000
137,800
57
Spring
331,100
327,600
328,000
87
Summer
158,900
158,700
159,200
44
Winter
290,600
288,100
289,100
94
To assess whether the monthly nitrate loading under all four scenarios (Baseline, 12%
(scenario 2) Cover Crops, and 24% (scenario 3) Cover Crops) at the Pamlico Estuary significantly
differs, we conducted an Analysis of Variance (ANOVA) (Table 3b.7). For cover crop application,
the ANOVA test yielded a p-value of 0.991, significantly greater than the significance level of 0.05.
Chapter 3b
157
This result indicates that we cannot reject the null hypothesis, suggesting there is not enough
statistical evidence to assert that the average monthly nitrate loading under the three scenarios
(baseline, 12% cover crops (scenario 2), and 24% (scenario 3) cover crops) at the Pamlico Estuary
are significantly different from each other (Table 3b.7).
Table 3b.7 Summary of ANOVA results for cover crop scenarios
ANOVA: Single Factor
SUMMARY
Groups
Count
Sum
Average
Variance
Baseline
204
29923200
146682.4
1.36E+10
Scenario-2
204
30140350
147746.8
1.37E+10
Scenario-3
204
30218900
148131.9
1.37E+10
Source of
Variation
SS
df
MS
F
p-value
F crit
Between
Groups
2.3E+08
2
1.15E+08
0.008407
0.991628
3.010517
Within
Groups
8.33E+12
609
1.37E+10
Total
8.33E+12
611
Chapter 3b
158
While the ANOVA test identifies overall differences across groups, it does not specify which
groups, if any, differ from one another (St & Wold, 1989). However, the high p-value suggests that
any differences in mean monthly nitrate loading among the various cover crop scenarios are likely
not significant. This is further underscored by the detailed summary and ANOVA table (Table
3b.7), which provide counts, sums, averages, and variances for each scenario. Despite minor
variations in average nitrate loads, these do not amount to statistically significant differences, as
evidenced by the Between Groups sum of squares (SS) of 230004436, a within-groups SS
exceeding 8.3 trillion, and a notably small F-value of 0.00841 against an F critical value of 3.0105.
This comprehensive analysis highlights that the variance observed within groups significantly
overshadows any minor differences between the groups (St & Wold, 1989). Such a conclusion
might indicate that the scale of cover crop implementation explored in the scenarios is insufficient
to establish a statistically significant impact on nitrate loading or winter wheat as a cover crop may
not be efficient in reducing nitrate loss from agricultural fields (Kaspar et al., 2012; Ritter et al.,
1998). It could also that the effectiveness of cover crop strategies might be influenced by factors
such as the spatial distribution of runoff sources, the timing of cover crop planting, and
management practices, as well as the presence of other nitrogen sources beyond agricultural runoff
(Kaspar et al., 2012).
The introduction of winter wheat as a cover crop could elevate nitrate loads due to several
interconnected factors (Nouri et al., 2022). The growth cycle of winter wheat, particularly its
dormant phase during early winter, may not align well with the timing of nitrate assimilation and
soil nitrogen release (Anthoni et al., 2004; Footitt et al., 2014). This misalignment can lead to
periods where nitrate, unutilized by the cover crop, becomes prone to leaching during the region's
heavy rainfall events, thereby contributing to increased nitrate loads in the watershed (Cordovil et
Chapter 3b
159
al., 2020; Robertson & Vitousek, 2009). This issue is compounded by the decomposition of winter
wheat post-harvest, which can release nitrogen back into the soil, subsequently increasing loss to
downstream waters. Furthermore, winter wheat's efficiency in nitrogen uptake might be
compromised by environmental conditions and management practices, leaving unused nitrate to
leach into the Tar-Pamlico basin. Soil disturbance from the planting and termination of winter
wheat exacerbates this problem by disrupting soil structure and potentially enhancing nitrate
mobility (McGuire et al., 1998; Moore et al., 2019).
This analysis clearly demonstrates that the relationship between cover crop coverage and
nitrate load is multifaceted, reflecting the influence of environmental conditions, management
practices, and the inherent characteristics of the watershed (Bindraban et al., 2012; Neumann et
al., 2021; Smith et al., 2019). It suggests that even while the literature suggests cover crops are
efficient in nutrient (nitrate) management strategies, their effectiveness is conditional,
necessitating a tailored approach based on watershed characteristics (Bindraban et al., 2012;
Neumann et al., 2021). This calls for continued research to identify the optimal conditions under
which cover crops can most effectively contribute to water quality improvement in agricultural
landscapes, underscoring the importance of an integrated approach to nutrient management.
3.4. Hydrological and Socio-Hydrological Model Comparison
The model simulated minimal variation in nitrate load at the outlet point for multiple BMP
scenarios, as detailed in section 3.3. Consequently, we conducted a comparison of Hydrological
and Socio-Hydrological models at the field scale. In SWAT+, each field is contained in an HRU
with fields of the same crop, management practices, soil type, and slope. Therefore, we compared
the nitrate levels leaving the agricultural fields under seven different scenarios at the HRU scale,
as shown in Table 3b.8. In this section, we present the economic and environmental implications
Chapter 3b
160
of cover crop and reduced fertilizer application rate conservation contracts through the use of the
hydrological model with naive assumptions of contract adoption (i.e. typical engineering
approach) and the Socio-Hydrological model with adoption informed by the farmer behavioral
model.
Chapter 3b
161
Table 3b.8 Cost-benefit analysis of different engineering and econometric scenarios
Policy
Scenarios
Additional
Area
Converted
(km2)
Total Budget Spent in
million dollars ($)
Area Enrolled
(km2)
Total annual
average
NO3-N (kg)
from
agriculture
Change total
annual
average
NO3-N (kg)
from
agriculture
Change in
nitrate per
enrolled area
(kg km-2)
Benefit per
enrolled area
($ km-2)
Benefit-cost
ratio
Cover crop ($75 per
acre or 18500 per km2)
Strict N application ($50
per Acre or $12300 per
km2)
$20.9 per kg
(Ribaudo et
al., 2005)
Scenario 1
0
0
0
1,013,000
Scenario 2
606
11.2
606
958,000
55,000
91.3
1,907
0.103
Scenario 3
1,212
22.4
1,212
957,000
56,000
46.5
971
0.053
Scenario 4
606
7.4
606
968,000
46,000
75.2
1,571
0.128
Scenario 5
1,212
15
1,212
940,000
73,000
60.5
1,265
0.103
Scenario 6
206
3.8
319
1,024,000
-11,000
-33
-700
-0.038
Scenario 7
46
0.6
46
1,009,000
4,000
97.3
2,033
0.165
Chapter 3b
162
In the context of the hydrological modeling scenarios detailed in the analysis, the 12% and
24% coverage figures represent the total percentage of agricultural land targeted for intervention,
assuming that all the farmland in these areas will agree to the offered conservation contract and
implement the specified land management practices. This assumption underpins the theoretical
maximum impact of the agricultural practices under consideration but may not align with real-
world applications. We used farmers’ behavioral model results (section 3.2) to develop the socio-
hydrological model. In the socio-hydrological model, 3.75% (reduced fertilizer application rate),
14.5% (additional cover crops applied), and 26.2% (total farmland enrolled for cover crops) reflect
the actual proportion of farmers willing to adopt these specific practices, based on informed
econometric analysis under standard EQIP cost-share rates.
The engineering approach (scenarios 2 through 5) assumed these areas would be targeted and
all the farmers in this area would implement the conservation practices. Assuming that all farmers
in a targeted area will adopt conservation practices can be overly optimistic, as it overlooks
individual differences in economic circumstances, willingness, and ability to change (Howard et
al., 2024). Integrating a farmers' behavioral model into the planning process can provide more
realistic predictions by accounting for these variations in behavior and receptivity, thereby
enhancing the effectiveness and efficiency of policy implementation (Howard et al., 2024). Thus,
for the socio-hydrological model implementation, these percentages are calculated relative to the
24% of agricultural land targeted in the hydrological scenarios. This means 3.75%, 14.5%, and
26.2% farmers of 24% of total selected agricultural HRUs will implement these practices on their
fields for reduced fertilizer application rate, additional cover crops applied, and total farmland
enrolled for cover crops, respectively. In other words, of the targeted farmers (24%), only 3.75%
would implement reduced fertilizer application making the practice implemented on 0.9% of the
Chapter 3b
163
agricultural land in the watershed. This differentiation in percentage base points—from theoretical
maximum to likely practical application—highlights the critical need to consider both the absolute
and relative scales when assessing the adoption and effectiveness of various agricultural and
environmental practices. The socio-hydrological modeling thus provides a more realistic estimate
of the area that would actually be converted to the new practices, bridging the gap between
theoretical planning and actual implementation (Howard et al., 2024).
3.4.1 Area Conversion and Budget Allocation:
The baseline scenario operates with no additional cost and no area conversion, serving as a control
for comparing Hydrological and Socio-Hydrological modeling scenarios for cover crops and
reduced fertilizer application rate (Table 3b.2).
Introducing cover crops on 12% of the agricultural lands in the watershed entails an
expenditure of approximately $11.2 million (considering the current EQIP rate for cover crops of
$185.3 ha-1) and affects over 606 km2. Doubling this coverage to 24% doubles the budget to about
$22.5 million and covers over 1200 km2. In contrast, a strategy focusing on a 30% reduction in
nitrogen usage over a similar area demands a lesser budget of around $7.5 million (considering the
current EQIP rate for strict N application of $123.5 ha-1). Based on the enrollment projections from
the socio-hydrological modeling scenarios involving cover crops and strict nitrogen (N)
application under current EQIP rates, the area converted within the targeted agricultural lands
amounts to over 206 km² for cover crops and over 45 km² for nitrogen application.
Correspondingly, the budgets allocated are more modest $3.8 million for cover crops and $0.6
million for nitrogen application.
Chapter 3b
164
3.4.2. Agricultural Output and Environmental Impact:
To examine the effect of the practices at the field scale we summed the cumulative annual average
agricultural nitrate loss (including surface, lateral, and groundwater nitrate) in kg from the
agricultural areas (HRU scale) in the Tar-Pamlico basin where the practices were implemented. To
compare the hydrological and socio-hydrological models, we considered only the 24% agricultural
area (as used in scenarios 3 and 5) to determine the cumulative annual average nitrate loss from
agricultural fields in all scenarios as mentioned in section 2.4.
The baseline scenario maintains agricultural nitrate-nitrogen export at approximately
1,010,000 kg without any interventions, serving as a reference for comparing various conservation
strategies. Introducing cover crops across 12% and 24% of the agricultural areas (Scenarios 2 and
3) reduces output slightly to 958,000 kg and 957,000 kg, respectively, which is a reduction of
55,300 and 56,300 kg of nitrate-nitrogen respectively. The extension of these strategies
demonstrates the environmental impact of scaling up sustainable practices. In Scenarios 4 and 5,
strict fertilizer reduction strategies applied over 12% and 24% of the agricultural areas further
influence export, which decreases to 967,800 kg and 935,500 kg, respectively. These changes
result in nitrate-nitrogen reductions of 45,600 kg and 77,900 kg, showing the positive
environmental effects of reduced nitrogen usage across varying scales. These changes highlight
the tradeoff in the farmers’ decision between decreased agricultural output, improved water quality,
and additional payment as part of the agricultural assistance program.
In Scenario 6, cover crops were applied to an additional 207 km2, bringing the total area under
this practice to 319 km2 square meters as part of the Environmental Quality Incentives Program
(EQIP). This expansion was guided by a farmers' behavioral model, which suggested practical
implementations based on actual farming practices. Implementing cover crops over this extended
Chapter 3b
165
area resulted in agricultural nitrate-nitrogen export of 1,020,100 kg and a slight increase in nitrate
levels by 10,600 kg. This scenario highlighted that even if the theoretical implementation of cover
crops reduced nitrogen load by targeting specific agricultural fields (HRU), it might not have been
possible in the real world. Engineering models specifically target HRUs that exhibit higher surface
nitrate exports, predicting that cover crops in these areas would effectively reduce nitrate loads
(scenario 2 & 3). However, when incorporating farmers' behavioral models, the selection of HRUs
isn't solely based on nitrate export levels, which means cover crops might be applied to fields
where they are less effective at reducing nitrates, as was observed in scenario 6. The effectiveness
of winter wheat when used as cover crops varied based on the location it was applied to, which
crops were planted on that field as the main summer cash crop, soil conditions, rainfall, elevation,
and other characteristics (Kaspar et al., 2012; Ritter et al., 1998). This discrepancy highlights the
potential limitations of applying naive hydrological models without considering the practical and
behavioral dynamics of the farmers.
In Scenario 7, a strict fertilizer application strategy is employed over an area representing
3.75% of the land managed by farmers willing to adopt this practice, as informed by a behavioral
model. This approach reduces the agricultural nitrate loss to 1,009,000 kg by decreasing the nitrate
load by 4,400 kg. This demonstrates the positive environmental impact of reducing nitrogen usage.
3.4.3 Economic Efficiency:
The benefit-cost analysis across the seven scenarios illustrates varying degrees of economic
efficiency, highlighting the effectiveness and cost implications of each agricultural strategy
implemented (Table 3b.8). To calculate the benefits for each scenario, we multiply the reduction
in nitrogen (N) by the estimated social damages of $20.90 for each additional kilogram of N
leaving the farm (Ribaudo et al., 2005).
Chapter 3b
166
In the baseline scenario (Scenario 1), no interventions are implemented, thus there is no
benefit-cost ratio, serving as a control for comparative purposes. For scenarios involving cover
crops, Scenario 2 (12% Cover Crop Application) achieves a benefit-cost ratio of 0.103, indicating
a moderate return on investment. This efficiency drops significantly in Scenario 3 (24% Cover
Crop Application) to a ratio of 0.052, demonstrating declining returns as the scale of cover crop
application increases. This result was expected, as we targeted HRUs in sequence according to
their nitrate loss in the hydrological modeling scenarios.
In terms of nitrogen management, Scenario 4 (30% Nitrogen Reduction over 12% Area)
shows the highest benefit-cost ratio among the initial assessments at 0.128, suggesting that targeted
nitrogen reduction is environmentally beneficial. Expanding this approach to a larger area in
Scenario 5 (30% Nitrogen Reduction over 24% Area) yields a lower ratio of 0.1028, indicating
reduced economic efficiency at a larger scale because of targeted HRU practice implementation in
hydrological modeling scenarios.
Scenario 7 (Strict Fertilizer Application over 3.75% Area), informed by a model of farmer
behavior, shows a modest increase in the benefit-cost ratio to 0.165. However, this rise from a
previous low of 0.05 is still considered suboptimal. More significantly, the disparity between the
hydrological model's predicted benefits for a 24% nitrogen reduction (65,300 kg) and the much
lower actual benefits observed in the socio-hydrological model (4,400 kg), which amount to only
about 6.8% of the expected reduction, suggests that the hydrological model may greatly
overestimate the efficacy of the policy. This overestimation might stem from the model's
assumption that farmers uniformly adopt and effectively implement BMPs, whereas in reality,
achieving targeted intervention could be challenging and costly, with varying levels of adoption
among farmers (Howard et al., 2024). This analysis confirms that while larger-scale interventions
Chapter 3b
167
may achieve significant environmental impacts, their economic viability can vary, often requiring
more nuanced, targeted approaches to maximize both ecological and economic benefits in
agricultural policymaking.
4. Conclusions
This study offers significant insights into the complexities of managing nutrient loading in coastal
watersheds through the integration of socio-hydrological modeling with farmers' behavioral
responses. The innovative approach of combining economic and engineering frameworks to
address accelerated eutrophication reveals several key findings from the study, offering important
considerations for policy and practice.
Socio-economic Dynamics of Agricultural Practice Adoption: Integrating farmers' behavioral
models into hydrological modeling efforts provided valuable insights into the socio-economic
dynamics influencing the adoption of sustainable agricultural practices. The study demonstrated
the importance (and limitations) of economic incentives in promoting BMP adoption among
farmers, as evidenced by the farmers' willingness to accept (WTA) different practices and the
highly significant impact of payment level on contract acceptance. However, the relatively low
willingness to adopt strict nitrogen application limits, even with incentives, points to the challenges
of changing farming practices and the need for policies that are both economically viable for
farmers and effective in environmental protection.
Limited Impact of Best Management Practices (BMPs) on Nitrate Loading: The simulation
results using the SWAT+ model show that, under the scenarios modeled, both cover crops and
reduced nitrogen application rates did not significantly reduce nitrate loading to the Pamlico
Chapter 3b
168
Estuary. These outcomes highlight the challenge of achieving measurable improvements in water
quality through isolated BMPs at the watershed scale, suggesting that the scale and scope of the
BMPs modeled may be insufficient for significant nitrate load reduction.
Hydrological Models Overestimate Policy Benefits: Our approach to implementing BMPs in
the hydrological models significantly overestimate the benefits of nitrogen reduction policies,
predicting a reduction that is 15 times higher than what is actually achieved—a discrepancy where
only 6.8% of the expected benefits are realized. This highlights the critical need for models to
integrate socio-economic dynamics and farmer behaviors to enhance the accuracy and
effectiveness of agricultural environmental policies.
Identifying diverse sources of nitrate in coastal watersheds: Our findings reveal the complexity
of nutrient transport, suggesting that agricultural practices may not be the main source of nitrate
in coastal watersheds. Effective nutrient management calls for a multifaceted approach that goes
beyond traditional agricultural BMPs, incorporating considerations of alternative sources such as
groundwater and legacy nitrate.
Importance of Interdisciplinary Approaches: This research underscores the value of
interdisciplinary approaches in tackling complex environmental issues like accelerated
eutrophication. By bridging econometric and engineering disciplines, the study provides a
comprehensive perspective on the challenges and opportunities for implementing sustainable
agricultural practices, emphasizing the need for collaboration across scientific, policy, and farming
communities to develop and implement effective strategies.
Future Directions for Research and Policy: The study highlights several areas for further
research, including the exploration of alternative BMPs, the integration of additional socio-
Chapter 3b
169
economic factors into modeling efforts, and the examination of the cumulative effects of combined
BMPs at larger scales. Policymakers are encouraged to consider these findings in the development
of agricultural and environmental policies, recognizing the critical role of economic incentives,
farmer engagement, and integrated management strategies in improving ecosystem health.
Lastly, while the study's findings indicate that the modeled BMPs alone may not significantly
reduce nitrate loading in the Pamlico Estuary, they importantly illuminate the complex interplay
between agricultural practices, watershed hydrology, and ecosystem health. These insights
contribute valuable knowledge towards developing more effective strategies for combating
eutrophication and enhancing the sustainability of agricultural landscapes.
Appendix 1: https://github.com/EtheridgeLab/Tar_Pam_SWAT/blob/main/Appendix1.docx
Acknowledgements:
This research was made possible with support from the Center for Sustainable Energy and
Environmental Engineering and the Integrated Coastal Sciences Program at East Carolina
University. Funding was provided by the National Science Foundation [Grant Numbers: 2009185
and 2052889] and the Environmental Protection Agency [Grant Number: R840181].
Data Availability:
The data supporting the conclusions of this research are available upon reasonable request from
the corresponding author. The majority of the data utilized in this study is open source.
Chapter 3b
170
LLM Statement:
We utilized ChatGPT to enhance the grammar and sentence structure. Post-editing, we thoroughly
reviewed the entire document for accuracy.
References:
Anthoni, P. M., Freibauer, A., Kolle, O., & Schulze, E. D. (2004). Winter wheat carbon exchange
in Thuringia, Germany. Agricultural and Forest Meteorology, 121(1-2), 55-67.
Bai, T., Li, L., Mu, P. F., Pan, B. Z., & Liu, J. (2023). Impact of climate change on water transfer
scale of inter-basin water diversion project. Water Resources Management, 37(6), 2505-2525.
Barker, T., Hatton, K., O'Connor, M., Connor, L., & Moss, B. (2008). Effects of nitrate load on
submerged plant biomass and species richness: results of a mesocosm experiment.
Fundamental and Applied Limnology, 173(2), 89.
Behboudian, M., Anamaghi, S., Mahjouri, N., & Kerachian, R. (2023). Enhancing the resilience
of ecosystem services under extreme events in socio-hydrological systems: a spatio-temporal
analysis. Journal of Cleaner Production, 397, 136437.
Bindraban, P. S., van der Velde, M., Ye, L., Van den Berg, M., Materechera, S., Kiba, D. I., ... &
van Lynden, G. (2012). Assessing the impact of soil degradation on food production. Current
Opinion in Environmental Sustainability, 4(5), 478-488.
Birgand, F., Appelboom, T. W., Chescheir, G. M., & Skaggs, R. W. (2011). Estimating nitrogen,
phosphorus, and carbon fluxes in forested and mixed-use watersheds of the lower coastal plain
Chapter 3b
171
of North Carolina: Uncertainties associated with infrequent sampling. Transactions of the
ASABE, 54(6), 2099-2110.
Bruland, G. L., Hanchey, M. F., & Richardson, C. J. (2003). Effects of agriculture and wetland
restoration on hydrology, soils, and water quality of a Carolina bay complex. Wetlands Ecology
and Management, 11, 141-156.
Burnett, E. A., Wilson, R. S., Roe, B., Howard, G., Irwin, E., Zhang, W., & Martin, J. (2015).
Farmers, phosphorus and water quality: Part II. A descriptive report of beliefs, attitudes and
best management practices in the Maumee watershed of the western Lake Erie basin. The Ohio
State University, School of Environment & Natural Resources, Columbus, OH.
Celikkol Erbas, B., & Guven Solakoglu, E. (2017). In the presence of climate change, the use of
fertilizers and the effect of income on agricultural emissions. Sustainability, 9(11), 1989.
Chang, J., Wang, Y., Istanbulluoglu, E., Bai, T., Huang, Q., Yang, D., & Huang, S. (2015). Impact
of climate change and human activities on runoff in the Weihe River Basin, China. Quaternary
International, 380, 169-179.
Cheng, G., & Li, X. (2015). Integrated research methods in watershed science. Science China
Earth Sciences, 58, 1159-1168.
Cherry, K. A., Shepherd, M., Withers, P. J. A., & Mooney, S. J. (2008). Assessing the effectiveness
of actions to mitigate nutrient loss from agriculture: A review of methods. Science of the total
environment, 406(1-2), 1-23.
Chapter 3b
172
Coopersmith, E. J., Minsker, B. S., & Sivapalan, M. (2014). Patterns of regional hydroclimatic
shifts: An analysis of changing hydrologic regimes. Water Resources Research, 50(3), 1960-
1983.
Cordovil, Cláudia MdS, Shabtai Bittman, Luis M. Brito, Michael J. Goss, Derek Hunt, João Serra,
Cameron Gourley et al. (2020) "Climate-resilient and smart agricultural management tools to
cope with climate change-induced soil quality decline." In Climate change and soil
interactions, pp. 613-662. Elsevier,
Craswell, E. (2021). Fertilizers and nitrate pollution of surface and ground water: an increasingly
pervasive global problem. SN Applied Sciences, 3(4), 518.
Cross, T. K., & Summerfelt, R. C. (1986). BOD dynamics in small eutrophic lakes in relation to
artificial mixing. Lake and Reservoir Management, 2(1), 286-292.
Dabney, S. M., Delgado, J. A., & Reeves, D. W. (2001). Using winter cover crops to improve soil
and water quality. Communications in Soil Science and Plant Analysis, 32(7-8), 1221-1250.
De Bruijn, J. A., Smilovic, M., Burek, P., Guillaumot, L., Wada, Y., & Aerts, J. C. (2023). GEB v0.
1: a large-scale agent-based socio-hydrological model–simulating 10 million individual
farming households in a fully distributed hydrological model. Geoscientific Model
Development, 16(9), 2437-2454.
Dillman, D. A., Smyth, J. D., & Christian, L. M. (2014). Internet, phone, mail, and mixed-mode
surveys: The tailored design method. John Wiley & Sons.
Du, E., Tian, Y., Cai, X., Zheng, Y., Li, X., & Zheng, C. (2020). Exploring spatial heterogeneity
and temporal dynamics of human-hydrological interactions in large river basins with intensive
Chapter 3b
173
agriculture: A tightly coupled, fully integrated modeling approach. Journal of Hydrology, 591,
125313.
Dukes, M. D., & Evans, R. O. (2006). Impact of agriculture on water quality in the North Carolina
Middle Coastal Plain. Journal of irrigation and drainage engineering, 132(3), 250-262.
Footitt, S., Clay, H. A., Dent, K., & Finch‐Savage, W. E. (2014). Environment sensing in spring‐
dispersed seeds of a winter annual A rabidopsis influences the regulation of dormancy to align
germination potential with seasonal changes. New Phytologist, 202(3), 929-939.
Gaur, V., Lang, C., Howard, G., & Quainoo, R. (2023). When energy issues are land use issues:
estimating preferences for utility-scale solar energy siting. Land Economics, 99(3), 343-363.
Genova, P., & Wei, Y. (2023). A socio-hydrological model for assessing water resource allocation
and water environmental regulations in the Maipo River basin. Journal of Hydrology, 617,
129159.
Gibson, C. A., Meyer, J. L., Poff, N. L., Hay, L. E., & Georgakakos, A. (2005). Flow regime
alterations under changing climate in two river basins: implications for freshwater ecosystems.
River Research and Applications, 21(8), 849-864.
Habib, K. M. N. (2011). A random utility maximization (RUM) based dynamic activity scheduling
model: Application in weekend activity scheduling. Transportation, 38(1), 123-151.
Hautier, Y., Niklaus, P. A., & Hector, A. (2009). Competition for light causes plant biodiversity
loss after eutrophication. Science, 324(5927), 636-638.
Hawley, M. E., Jackson, T. J., & McCuen, R. H. (1983). Surface soil moisture variation on small
agricultural watersheds. Journal of Hydrology, 62(1-4), 179-200.
Chapter 3b
174
Hettiarachchi, S., Wasko, C., & Sharma, A. (2018). Increase in flood risk resulting from climate
change in a developed urban watershed–the role of storm temporal patterns. Hydrology and
Earth System Sciences, 22(3), 2041-2056.
Höök, M., & Tang, X. (2013). Depletion of fossil fuels and anthropogenic climate change—A
review. Energy policy, 52, 797-809.
Howard, G., & Roe, B. E. (2013). Stripping because you want to versus stripping because the
money is good: a latent class analysis of farmer preferences regarding filter strip programs.
Howard, G., Zhang, W., Valcu‐Lisman, A., & Gassman, P. W. (2024). Evaluating the tradeoff
between cost effectiveness and participation in agricultural conservation programs. American
Journal of Agricultural Economics, 106(2), 712-738.
Howarth, R., Chan, F., Conley, D. J., Garnier, J., Doney, S. C., Marino, R., & Billen, G. (2011).
Coupled biogeochemical cycles: eutrophication and hypoxia in temperate estuaries and coastal
marine ecosystems. Frontiers in Ecology and the Environment, 9(1), 18-26.
Kapsalis, V. C., & Kalavrouziotis, I. K. (2021). Eutrophication—A worldwide water quality issue.
Chemical Lake Restoration: Technologies, Innovations and Economic Perspectives, 1-21.
Kaspar, T. C., Jaynes, D. B., Parkin, T. B., Moorman, T. B., & Singer, J. W. (2012). Effectiveness
of oat and rye cover crops in reducing nitrate losses in drainage water. Agricultural Water
Management, 110, 25-33.
Kuil, L., Evans, T., McCord, P. F., Salinas, J. L., & Blöschl, G. (2018). Exploring the influence of
smallholders' perceptions regarding water availability on crop choice and water allocation
through socio‐hydrological modeling. Water Resources Research, 54(4), 2580-2604.
Chapter 3b
175
Lal, R. (2013). Intensive agriculture and the soil carbon pool. Journal of Crop Improvement, 27(6),
735-751.
Liu, T., Bruins, R. J., & Heberling, M. T. (2018). Factors influencing farmers’ adoption of best
management practices: A review and synthesis. Sustainability, 10(2), 432.
Liu, Y., Gupta, H., Springer, E., & Wagener, T. (2008). Linking science with environmental
decision making: Experiences from an integrated modeling approach to supporting sustainable
water resources management. Environmental Modelling & Software, 23(7), 846-858.
Luo, Y., Ficklin, D. L., Liu, X., & Zhang, M. (2013). Assessment of climate change impacts on
hydrology and water quality with a watershed modeling approach. Science of the total
environment, 450, 72-82.
Ma, S., Swinton, S. M., Lupi, F., & Jolejole‐Foreman, C. (2012). Farmers’ willingness to
participate in Payment‐for‐Environmental‐Services programmes. Journal of Agricultural
Economics, 63(3), 604-626.
McGuire, A. M., Bryant, D. C., & Denison, R. F. (1998). Wheat yields, nitrogen uptake, and soil
moisture following winter legume cover crop vs. fallow. Agronomy journal, 90(3), 404-410.
Moore, Kenneth J., Robert P. Anex, Amani E. Elobeid, Shuizhang Fei, Cornelia B. Flora, A. Susana
Goggi, Keri L. Jacobs et al. (2019) "Regenerating agricultural landscapes with perennial
groundcover for intensive crop production." Agronomy 9, no. 8 (2019): 458.
NADP National Atmospheric Deposition Program. (2023). Retrieved December, 2023, from
https://nadp.slh.wisc.edu/
Chapter 3b
176
Naganur, S., Patil, N. S., Patil, V., & Pujar, G. (2024). Evaluation of best management practices
(BMPS) and their impact on environmental flow through SWAT+ model. Modeling Earth
Systems and Environment, 1-15.
NCDEQ North Carolina Department of Environmental Quality. (2014). Tar-Pamlico Basin Plan.
North Carolina Department of Environmental Quality.
https://www.deq.nc.gov/about/divisions/water-resources/water-planning/basin-
planning/river-basin-plans/tar-pamlico#2014Tar-PamlicoBasinPlan-4040
NCGAR North Carolina Department of Agriculture and Consumer Services. (2022). Tar-Pamlico
Watershed Initiatives. North Carolina Department of Agriculture and Consumer Services.
https://www.ncagr.gov/divisions/soil-water-conservation/programs-initiatives/watershed-
initiatives/tar-pamlico#Contact-4304
Neumann, A., Dong, F., Shimoda, Y., Arnillas, C. A., Javed, A., Yang, C., ... & Arhonditsis, G. B.
(2021). A review of the current state of process-based and data-driven modelling: guidelines
for Lake Erie managers and watershed modellers. Environmental Reviews, 29(4), 443-490.
Nolan, B. T., Healy, R. W., Taber, P. E., Perkins, K., Hitt, K. J., & Wolock, D. M. (2007). Factors
influencing ground-water recharge in the eastern United States. Journal of Hydrology, 332(1-
2), 187-205.
Nouri, A., Lukas, S., Singh, S., Singh, S., & Machado, S. (2022). When do cover crops reduce
nitrate leaching? A global meta‐analysis. Global Change Biology, 28(15), 4736-4749.
Ouyang, Y., Grace, J. M., Parajuli, P. B., & Caldwell, P. V. (2022). Impacts of multiple hurricanes
and tropical storms on watershed hydrological processes in the Florida panhandle. Climate,
10(3), 42.
Chapter 3b
177
Palmer, M. A., Reidy Liermann, C. A., Nilsson, C., Flörke, M., Alcamo, J., Lake, P. S., & Bond,
N. (2008). Climate change and the world's river basins: anticipating management options.
Frontiers in Ecology and the Environment, 6(2), 81-89.
Peterson, J. M., Smith, C. M., Leatherman, J. C., Hendricks, N. P., & Fox, J. A. (2015). Transaction
costs in payment for environmental service contracts. American Journal of Agricultural
Economics, 97(1), 219-238.
Peterson, T. J., Saft, M., Peel, M. C., & John, A. (2021). Watersheds may not recover from drought.
Science, 372(6543), 745-749.
Petry, J., Soulsby, C., Malcolm, I. A., & Youngson, A. F. (2002). Hydrological controls on nutrient
concentrations and fluxes in agricultural catchments. Science of the Total Environment, 294(1-
3), 95-110.
Reckhow, K. H., & Chapra, S. C. (1999). Modeling excessive nutrient loading in the environment.
Environmental Pollution, 100(1-3), 197-207.
Ribaudo, M. O., R. Heimlich, and M. Peters. 2005. “Nitrogen Sources and Gulf Hypoxia: Potential
for Environmental Credit Trading.” Ecological Economics 52(2): 159–68.
Ritter, W. F., Scarborough, R. W., & Chirnside, A. E. M. (1998). Winter cover crops as a best
management practice for reducing nitrogen leaching. Journal of Contaminant Hydrology, 34(1-
2), 1-15.
Robertson, G. P., & Vitousek, P. M. (2009). Nitrogen in agriculture: balancing the cost of an
essential resource. Annual review of environment and resources, 34, 97-125.
Chapter 3b
178
Rossi, A. M., & Hinrichs, C. C. (2011). Hope and skepticism: Farmer and local community views
on the socio-economic benefits of agricultural bioenergy. Biomass and Bioenergy, 35(4), 1418-
1428.
Schuerz, C. (2019). chrisschuerz/SWATplusR: SWATplusR 0.2. 7.
Sheshukov, A. Y., Douglas-Mankin, K. R., Sinnathamby, S., & Daggupati, P. (2016). Pasture BMP
effectiveness using an HRU-based subarea approach in SWAT. Journal of environmental
management, 166, 276-284.
Shibu, M. E., Leffelaar, P. A., Van Keulen, H., & Aggarwal, P. K. (2010). LINTUL3, a simulation
model for nitrogen-limited situations: Application to rice. European Journal of Agronomy,
32(4), 255-271.
Singh, S.; Hwang, S.; Arnold, J.G.; Bhattarai, R. Evaluation of Agricultural BMPs’ Impact on
Water Quality and Crop Production Using SWAT+ Model. Agriculture 2023, 13, 1484.
https://doi.org/10.3390/agriculture13081484
Smith, D. R., Macrae, M. L., Kleinman, P. J. A., Jarvie, H. P., King, K. W., & Bryant, R. B. (2019).
The latitudes, attitudes, and platitudes of watershed phosphorus management in North
America. Journal of environmental quality, 48(5), 1176-1190.
St, L., & Wold, S. (1989). Analysis of variance (ANOVA). Chemometrics and intelligent
laboratory systems, 6(4), 259-272.
SWAT+ IO Document. (2020). Input/output file documentation, version 2016 modified on
November 16, 2020 according to REV 60.5. Retrieved from https://swatplus.gitbook.io/io-
docs
Chapter 3b
179
Upadhyay, P., Linhoss, A., Kelble, C., Ashby, S., Murphy, N., & Parajuli, P. B. (2022). Applications
of the SWAT model for coastal watersheds: review and recommendations. Journal of the
ASABE, 65(2), 453-469.
Veisi, H., Jackson-Smith, D., & Arrueta, L. (2022). Alignment of stakeholder and scientist
understandings and expectations in a participatory modeling project. Environmental Science
& Policy, 134, 57-66.
Wendling, M., Büchi, L., Amossé, C., Sinaj, S., Walter, A., & Charles, R. (2016). Influence of root
and leaf traits on the uptake of nutrients in cover crops. Plant and Soil, 409, 419-434.
Wurtsbaugh, W. A., Paerl, H. W., & Dodds, W. K. (2019). Nutrients, eutrophication and harmful
algal blooms along the freshwater to marine continuum. Wiley Interdisciplinary Reviews:
Water, 6(5), e1373.
