**Watershed-Scale Hydrological Modeling Methods and Applications**

Prem B. Parajuli and Ying Ouyang

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/53596

## **1. Introduction**

Pollution of surface water with harmful chemicals and eutrophication of rivers and lakes with excess nutrients are serious environmental concerns. The U.S. Environmental Protec‐ tion Agency (USEPA) estimated that 53% of the 27% assessed rivers and streams miles and 69% of the 45% assessed lakes, ponds, and reservoirs acreage in the nation are impaired (USEPA, 2010). In Mississippi, 57% of the 5% assessed rivers and streams miles are impaired (USEPA, 2010). These impairment estimates may increase when assessments of more water bodies are performed and water quality criteria are improved. The most common water pol‐ lution concerns in U.S. rivers and streams are sediment, nutrients (Phosphorus and Nitro‐ gen) and pathogens. Hydrological processes can significantly impact on the transport of water quality pollutants.

Non-point source pollution from agricultural, forest, and urban lands can contribute to wa‐ ter quality degradation. Total Maximum Daily Loads (TMDLs) are developed by states to improve water quality. The TMDL requires identifying and quantifying pollutant contribu‐ tions from each source to devise source-specific pollutant reduction strategies to meet appli‐ cable water quality standards. Commonly, water quality assessment at the watershed scale is accomplished using two techniques: (a) watershed monitoring and (b) watershed model‐ ing. Watershed models provide a tool for linking pollutants to the receiving streams. Models provide quick and cost-effective assessment of water quality conditions, as they can simu‐ late hydrologic processes, which are affected by several factors including climate change, soils, and agricultural management practices. However, methods used to develop a model for watersheds can significantly impact in the model outputs. Here several hydrological and water quality models are described. Case studies of two commonly used models with cali‐ bration and validation are provided with current and future climate change scenarios. This

© 2013 Parajuli and Ouyang; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Parajuli and Ouyang; licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

book chapter briefly reviews currently available hydrologic and water quality models, and presents model application case studies, to provide a foundation for further model develop‐ ment and watershed assessment studies.

Resources in Rural Basins (SWRRB) and Routing Outputs to Outlet (ROTO) models. The SWAT model development was influenced by other models like CREAMS (Knisel, 1980),

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

59

The SWAT model has been recently applied to assess watershed conditions of the U.S. (Gassman et al., 2007; Parajuli et al., 2008; 2009; Parajuli 2010a; 2011; 2012; Chaubey et al., 2010) and internationally such as Ethiopia (Betrie et al., 2011); Kenya and northwest Tanzania (Dessu and Melesse, 2012); Bulgaria and Greece (Boskidis et al., 2012); and

The AnnAGNPS model is a product of the USDA Agriculture Research Service (USDA-ARS) and the USDA Natural Resources Conservation Service (USDA-NRCS) to evaluate nonpoint source pollution from agriculture watersheds. Similar to the SWAT model, it is a phys‐ ically based continuous and daily time step model used to simulate surface runoff, sediment, and nutrient yields (Cronshey and Theurer, 1998; Bingner and Theurer, 2003). The AnnAGNPS is considered an enhanced modification to the single event based Agricultural Non-Point Source (AGNPS) model (Young et al., 1989), as it retains many features of AGNPS (Yuan et al., 2001). Unlike AGNPS, the AnnAGNPS delineates watershed, sub-di‐ vides the watershed into small drainage areas with homogenous land use, soils, etc. The sub-areas are integrated and simulated surface runoff and pollutant loads through rivers and streams within the sub-areas and watershed, which is enhanced from the AGNPS.

The AnnAGNPS model utilizes and incorporates components or sub-components from sev‐ eral other models such as; Revised Universal Soil Loss Equation (RUSLE) model (Renard et al., 1997); Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS) model (Knisel, 1980); Groundwater Loading Effects on Agricultural Manage‐ ment Systems (GLEAMS) model (Leonard et al., 1987); and Erosion Productivity Impact Cal‐ culator (EPIC) Model (Sharpley and Williams, 1990). The AnnAGNPS model represents small watershed areas using a cell-based approach, with land and soil property characteriza‐ tion similar to SWAT model HRUs. Daily soil moisture contents are calculated using the Curve Number (CN) method, which help to quantify surface and subsurface flows. The An‐

Refereed AnnAGNPS model based evaluations have been applied predominantly to water‐ sheds located in the U.S. (Yuan et al., 2011; 2002; Zuercher et al., 2011; Polyakov et al., 2007). However, the model also has been applied in other countries such as Mediterranian (Licciar‐

The Water Erosion Prediction Project (WEPP) model is a product of USDA. The WEPP model is a process-based, distributed parameter, single storm and continuous based model used to predict surface flow and sediment yields from the hill slopes and small watersheds. WEPP allows simulation of the effects of crop, crop rotation, contour farm‐

dello et al., 2011; 2007); Australia (Baginska et al., 2003), and China (Hua et al., 2012).

nAGNPS model uses the RUSLE to estimate sediment yields.

GLEAMS (Leonard et al., 1987), and EPIC (Williams et al., 1984; Neitsch et al., 2002).

Australia (Githui et al., 2012).

**2.2. AnnAGNPS**

**2.3. WEPP**

## **2. Review of water quality models**

Several useful hydrologic and water quality models are available today, each with diverse capabilities for watershed assessment. Many of these models are relevant to water quality goal assessment and implementation. Modeling of hydrology, sediment and nutrients has developed substantially, but advances have not always been consistent with the needs of the water quality goals program. Comprehensive education and training with model applica‐ tions and case studies are needed for users to understand the potentials, limitations, and suitable applications of a model. Review of several hydrological models (e.g. SWAT, An‐ nAGNPS, HSPF, SPARROW, GLEAMS, WEPP, EFDC etc.) including models description and application within the U.S. or other countries are discussed.

## **2.1. SWAT model**

The SWAT model is developed and supported by the USDA/ARS. It is a physically based watershed-scale continuous time-scale model, which operates on a daily time step. The SWAT model can simulate runoff, sediment, nutrients, pesticide, and bacteria transport from agricultural watersheds (Arnold et al., 1998). The SWAT model delineates a watershed, and sub-divides that watershed in to sub-basins. In each sub-basin, the model creates sever‐ al hydrologic response units (HRUs) based on specific land cover, soil, and topographic con‐ ditions. Model simulations that are performed at the HRU levels are summarized for the sub-basins. Water is routed from HRUs to associated reaches in the SWAT model. SWAT first deposits estimated pollutants within the stream channel system then transport them to the outlet of the watershed. The HRUs provide opportunity to include processes for possible spatial and temporal variations in model input parameters. The hydrologic module of the model quantifies a soil water balance at each time step during the simulation period based on daily precipitation inputs.

The SWAT model distinguishes the effects of weather, surface runoff, evapo-transpiration, crop growth, nutrient loading, water routing, and the long-term effects of varying agricul‐ tural management practices (Neitsch et al., 2005). In the hydrologic module of the model, the surface runoff is estimated separately for each sub-basin and routed to quantify the total surface runoff for the watershed. Runoff volume is commonly estimated from daily rainfall using modified SCS-CN method. The Modified Universal Soil Loss Equation (MUSLE) is used to predict sediment yield from the watershed. The SWAT model has been extensively applied for simulating stream flow, sediment yield, and nutrient modeling (Gosain et al., 2005; Vache et al., 2002; Varanou et al., 2002). The model needs several data inputs to repre‐ sent watershed conditions which include: digital elevation model (DEM), land use land cov‐ er, soils, climate data. The SWAT model is an advancement of the Simulator for Water Resources in Rural Basins (SWRRB) and Routing Outputs to Outlet (ROTO) models. The SWAT model development was influenced by other models like CREAMS (Knisel, 1980), GLEAMS (Leonard et al., 1987), and EPIC (Williams et al., 1984; Neitsch et al., 2002).

The SWAT model has been recently applied to assess watershed conditions of the U.S. (Gassman et al., 2007; Parajuli et al., 2008; 2009; Parajuli 2010a; 2011; 2012; Chaubey et al., 2010) and internationally such as Ethiopia (Betrie et al., 2011); Kenya and northwest Tanzania (Dessu and Melesse, 2012); Bulgaria and Greece (Boskidis et al., 2012); and Australia (Githui et al., 2012).

## **2.2. AnnAGNPS**

book chapter briefly reviews currently available hydrologic and water quality models, and presents model application case studies, to provide a foundation for further model develop‐

Several useful hydrologic and water quality models are available today, each with diverse capabilities for watershed assessment. Many of these models are relevant to water quality goal assessment and implementation. Modeling of hydrology, sediment and nutrients has developed substantially, but advances have not always been consistent with the needs of the water quality goals program. Comprehensive education and training with model applica‐ tions and case studies are needed for users to understand the potentials, limitations, and suitable applications of a model. Review of several hydrological models (e.g. SWAT, An‐ nAGNPS, HSPF, SPARROW, GLEAMS, WEPP, EFDC etc.) including models description

The SWAT model is developed and supported by the USDA/ARS. It is a physically based watershed-scale continuous time-scale model, which operates on a daily time step. The SWAT model can simulate runoff, sediment, nutrients, pesticide, and bacteria transport from agricultural watersheds (Arnold et al., 1998). The SWAT model delineates a watershed, and sub-divides that watershed in to sub-basins. In each sub-basin, the model creates sever‐ al hydrologic response units (HRUs) based on specific land cover, soil, and topographic con‐ ditions. Model simulations that are performed at the HRU levels are summarized for the sub-basins. Water is routed from HRUs to associated reaches in the SWAT model. SWAT first deposits estimated pollutants within the stream channel system then transport them to the outlet of the watershed. The HRUs provide opportunity to include processes for possible spatial and temporal variations in model input parameters. The hydrologic module of the model quantifies a soil water balance at each time step during the simulation period based

The SWAT model distinguishes the effects of weather, surface runoff, evapo-transpiration, crop growth, nutrient loading, water routing, and the long-term effects of varying agricul‐ tural management practices (Neitsch et al., 2005). In the hydrologic module of the model, the surface runoff is estimated separately for each sub-basin and routed to quantify the total surface runoff for the watershed. Runoff volume is commonly estimated from daily rainfall using modified SCS-CN method. The Modified Universal Soil Loss Equation (MUSLE) is used to predict sediment yield from the watershed. The SWAT model has been extensively applied for simulating stream flow, sediment yield, and nutrient modeling (Gosain et al., 2005; Vache et al., 2002; Varanou et al., 2002). The model needs several data inputs to repre‐ sent watershed conditions which include: digital elevation model (DEM), land use land cov‐ er, soils, climate data. The SWAT model is an advancement of the Simulator for Water

ment and watershed assessment studies.

**2. Review of water quality models**

**2.1. SWAT model**

on daily precipitation inputs.

and application within the U.S. or other countries are discussed.

58 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

The AnnAGNPS model is a product of the USDA Agriculture Research Service (USDA-ARS) and the USDA Natural Resources Conservation Service (USDA-NRCS) to evaluate nonpoint source pollution from agriculture watersheds. Similar to the SWAT model, it is a phys‐ ically based continuous and daily time step model used to simulate surface runoff, sediment, and nutrient yields (Cronshey and Theurer, 1998; Bingner and Theurer, 2003). The AnnAGNPS is considered an enhanced modification to the single event based Agricultural Non-Point Source (AGNPS) model (Young et al., 1989), as it retains many features of AGNPS (Yuan et al., 2001). Unlike AGNPS, the AnnAGNPS delineates watershed, sub-di‐ vides the watershed into small drainage areas with homogenous land use, soils, etc. The sub-areas are integrated and simulated surface runoff and pollutant loads through rivers and streams within the sub-areas and watershed, which is enhanced from the AGNPS.

The AnnAGNPS model utilizes and incorporates components or sub-components from sev‐ eral other models such as; Revised Universal Soil Loss Equation (RUSLE) model (Renard et al., 1997); Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS) model (Knisel, 1980); Groundwater Loading Effects on Agricultural Manage‐ ment Systems (GLEAMS) model (Leonard et al., 1987); and Erosion Productivity Impact Cal‐ culator (EPIC) Model (Sharpley and Williams, 1990). The AnnAGNPS model represents small watershed areas using a cell-based approach, with land and soil property characteriza‐ tion similar to SWAT model HRUs. Daily soil moisture contents are calculated using the Curve Number (CN) method, which help to quantify surface and subsurface flows. The An‐ nAGNPS model uses the RUSLE to estimate sediment yields.

Refereed AnnAGNPS model based evaluations have been applied predominantly to water‐ sheds located in the U.S. (Yuan et al., 2011; 2002; Zuercher et al., 2011; Polyakov et al., 2007). However, the model also has been applied in other countries such as Mediterranian (Licciar‐ dello et al., 2011; 2007); Australia (Baginska et al., 2003), and China (Hua et al., 2012).

#### **2.3. WEPP**

The Water Erosion Prediction Project (WEPP) model is a product of USDA. The WEPP model is a process-based, distributed parameter, single storm and continuous based model used to predict surface flow and sediment yields from the hill slopes and small watersheds. WEPP allows simulation of the effects of crop, crop rotation, contour farm‐ ing, and strip cropping. The WEPP model components includes weather generation, snow accumulation and melt, irrigation, infiltration, overland flow process, water bal‐ ance, plant growth, residue management, soil disturbance by tillage, and erosion process‐ es. The WEPP model considers sheet and rill erosion processes to predict erosion. The WEPP model incorporates modified water balance and percolation components from the SWRRB model (Williams and Nicks, 1985). The WEPP model utilizes and incorporates components or sub-components from several other models such as; EPIC (Williams et al., 1984); and CREAMS model (Knisel, 1980). The WEPP model has undergone continuous development since 1992 (1992-1995 with DOS version; 1997-2000 with window interface; 1999-2009 with Geo-WEPP ArcView/ArcGIS extensions; and 2001-present with webbrowser interface; Flanagan et al., 2007; Foltz et al., 2011).

HSPF model can simulate three sediment types (sand, silt, and clay) in addition to organic chemicals and alternative products. A detailed description of HSPF model can be found in

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

61

There have been hundreds of applications of HSPF around the world (Bicknell et al., 2001; Akter and Babel, 2012; Ouyang et al., 2012; Rolle et al., 2012). Examples include applications in a large watershed at the Chesapeake Bay, in a small watershed near Watkinsville, GA, with the experimental plots of a few hectares and in other areas such as Seattle, WA, Patux‐ ent River, MD., and Truckee-Carson Basins, NV. Details are available at: (http://

The SPAtially-Referenced Regression On Watershed attributes (SPARROW) model is a wa‐ tershed modeling tool for comparing water-quality data collected at a network of monitor‐ ing stations to characterize watersheds containing the stations (Smith et al., 1997; Schwarz et al., 2008). The SPARROW model has a nonlinear regression equation depicting the non-con‐ servative transport of contaminants from the point and diffuse sources on land surfaces to streams and rivers. The SPARROW predicts contaminant flux, concentration, and yield in streams. It has been used to evaluate alternative hypotheses about important contaminant sources and watershed properties that control contaminant load and transport over large spatial scales. The SPARROW can be used to explain spatial patterns of stream water quality

Numerous applications of SPARROW have been performed to assess water quality in wa‐ tersheds in recent years. Brown (2011) investigated nutrient sources and transport in the Missouri River Basin with SPARROW. Saad et al. (2011) applied SPARROW to estimate nu‐ trient load and to improve water quality monitoring design using a multi-agency dataset. Alam and Goodall (2012) examined the effects of hydrologic and nitrogen source changes on

The Environmental Fluid Dynamics Code (EFDC) is a multifunctional surface water model‐ ing system, which includes hydrodynamic, sediment-contaminant, and eutrophication com‐ ponents (Hamrick, 1996) and is available to the public through US-EPA website available at: http://www.epa.gov/ceampubl/swater/efdc/index.html. The EFDC can be used to simulate aquatic systems in multiple dimensions with the stretched or sigma vertical coordinates and the Cartesian (or curvilinear), and orthogonal horizontal coordinates to represent the physi‐ cal characteristics of a water body. A dynamically-coupled transport process for turbulent kinetic energy, turbulent length scale, salinity and temperature are included in the EFDC model. The EFDC allows for drying and wetting in shallow water bodies by a mass conser‐

Refereed EFDC-model-based evaluations exist predominately for stream ecosystems. Exam‐ ples include a three-dimensional hydrodynamic model of the Chicago River, Illinois (Sinha

Bicknell et al. (2001).

**2.6. SPARROW**

**2.7. EFDC**

vation scheme.

water.usgs.gov/cgi-bin/man\_wrdapp?hspf).

in relation to human activities and natural processes.

nitrogen yield in the contiguous United States with SPARROW.

Refereed WEPP-model-based evaluations exist predominantly for agricultural fields or small watersheds located in the U.S. (Dun et al., 2010; Flanagan et al., 2007; Foltz et al., 2011). However, the WEPP has been applied in other countries such as China (Zhang et al., 2008).

## **2.4. GLEAMS**

Groundwater Loading Effects of Agricultural Management Systems (GLEAMS) is a dai‐ ly time-step, continuous, field-scale hydrological and pollutant transport mathematical model (Leonard et al., 1987). The GLEAMS model can simulate surface runoff, percola‐ tion, nutrient and pesticide leaching, erosion and sedimentation. The GLEAMS model requires several daily climate data including mean daily air temperature, daily rainfall, mean monthly maximum and minimum temperatures, wind speed, solar radiation and dew-point temperature data. The soil input parameters in the model can be obtained from the State Soil Geographic Database (STATSGO) or Soil Survey Geographic Data‐ base (SSURGO) soil data. Previous studies described the ability of GLEAMS model to predict nitrate transport process from the agricultural areas (Shirmohammadi et al., 1998; Bakhsh et al., 2000; Chinkuyu and Kanwar, 2001).

Refereed GLEAMS model applications have been published predominantly for field scale studies in the U.S. (Bakhsh et al., 2000; Chinkuyu et al., 2004). However, GLEAMS also has been applied in a few other countries, such as China (Zhang et al., 2008).

## **2.5. HSPF model**

The hydrological simulation program—FORTRAN (HSPF) is a product of U.S. Environmen‐ tal Protection Agency (US-EPA), which is a comprehensive model used for modeling proc‐ esses related to water quantity and quality in watersheds of various sizes and complexities (Bicknell et al. 2001). It simulates both the land area of watersheds and the water bodies. The HSPF model uses input data including hourly history of rainfall, temperature and solar radi‐ ation; land surface characteristics/land use conditions; and land management practices to predict parameters at watershed scales. The results of model simulations are based on a time history of the quantity and quality of runoff from an urban, forest or agricultural watershed, which include surface runoff, sediment load, nutrients and pesticide concentrations. The HSPF model can simulate three sediment types (sand, silt, and clay) in addition to organic chemicals and alternative products. A detailed description of HSPF model can be found in Bicknell et al. (2001).

There have been hundreds of applications of HSPF around the world (Bicknell et al., 2001; Akter and Babel, 2012; Ouyang et al., 2012; Rolle et al., 2012). Examples include applications in a large watershed at the Chesapeake Bay, in a small watershed near Watkinsville, GA, with the experimental plots of a few hectares and in other areas such as Seattle, WA, Patux‐ ent River, MD., and Truckee-Carson Basins, NV. Details are available at: (http:// water.usgs.gov/cgi-bin/man\_wrdapp?hspf).

## **2.6. SPARROW**

ing, and strip cropping. The WEPP model components includes weather generation, snow accumulation and melt, irrigation, infiltration, overland flow process, water bal‐ ance, plant growth, residue management, soil disturbance by tillage, and erosion process‐ es. The WEPP model considers sheet and rill erosion processes to predict erosion. The WEPP model incorporates modified water balance and percolation components from the SWRRB model (Williams and Nicks, 1985). The WEPP model utilizes and incorporates components or sub-components from several other models such as; EPIC (Williams et al., 1984); and CREAMS model (Knisel, 1980). The WEPP model has undergone continuous development since 1992 (1992-1995 with DOS version; 1997-2000 with window interface; 1999-2009 with Geo-WEPP ArcView/ArcGIS extensions; and 2001-present with web-

Refereed WEPP-model-based evaluations exist predominantly for agricultural fields or small watersheds located in the U.S. (Dun et al., 2010; Flanagan et al., 2007; Foltz et al., 2011). However, the WEPP has been applied in other countries such as China (Zhang et al., 2008).

Groundwater Loading Effects of Agricultural Management Systems (GLEAMS) is a dai‐ ly time-step, continuous, field-scale hydrological and pollutant transport mathematical model (Leonard et al., 1987). The GLEAMS model can simulate surface runoff, percola‐ tion, nutrient and pesticide leaching, erosion and sedimentation. The GLEAMS model requires several daily climate data including mean daily air temperature, daily rainfall, mean monthly maximum and minimum temperatures, wind speed, solar radiation and dew-point temperature data. The soil input parameters in the model can be obtained from the State Soil Geographic Database (STATSGO) or Soil Survey Geographic Data‐ base (SSURGO) soil data. Previous studies described the ability of GLEAMS model to predict nitrate transport process from the agricultural areas (Shirmohammadi et al.,

Refereed GLEAMS model applications have been published predominantly for field scale studies in the U.S. (Bakhsh et al., 2000; Chinkuyu et al., 2004). However, GLEAMS also has

The hydrological simulation program—FORTRAN (HSPF) is a product of U.S. Environmen‐ tal Protection Agency (US-EPA), which is a comprehensive model used for modeling proc‐ esses related to water quantity and quality in watersheds of various sizes and complexities (Bicknell et al. 2001). It simulates both the land area of watersheds and the water bodies. The HSPF model uses input data including hourly history of rainfall, temperature and solar radi‐ ation; land surface characteristics/land use conditions; and land management practices to predict parameters at watershed scales. The results of model simulations are based on a time history of the quantity and quality of runoff from an urban, forest or agricultural watershed, which include surface runoff, sediment load, nutrients and pesticide concentrations. The

browser interface; Flanagan et al., 2007; Foltz et al., 2011).

60 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

1998; Bakhsh et al., 2000; Chinkuyu and Kanwar, 2001).

been applied in a few other countries, such as China (Zhang et al., 2008).

**2.4. GLEAMS**

**2.5. HSPF model**

The SPAtially-Referenced Regression On Watershed attributes (SPARROW) model is a wa‐ tershed modeling tool for comparing water-quality data collected at a network of monitor‐ ing stations to characterize watersheds containing the stations (Smith et al., 1997; Schwarz et al., 2008). The SPARROW model has a nonlinear regression equation depicting the non-con‐ servative transport of contaminants from the point and diffuse sources on land surfaces to streams and rivers. The SPARROW predicts contaminant flux, concentration, and yield in streams. It has been used to evaluate alternative hypotheses about important contaminant sources and watershed properties that control contaminant load and transport over large spatial scales. The SPARROW can be used to explain spatial patterns of stream water quality in relation to human activities and natural processes.

Numerous applications of SPARROW have been performed to assess water quality in wa‐ tersheds in recent years. Brown (2011) investigated nutrient sources and transport in the Missouri River Basin with SPARROW. Saad et al. (2011) applied SPARROW to estimate nu‐ trient load and to improve water quality monitoring design using a multi-agency dataset. Alam and Goodall (2012) examined the effects of hydrologic and nitrogen source changes on nitrogen yield in the contiguous United States with SPARROW.

## **2.7. EFDC**

The Environmental Fluid Dynamics Code (EFDC) is a multifunctional surface water model‐ ing system, which includes hydrodynamic, sediment-contaminant, and eutrophication com‐ ponents (Hamrick, 1996) and is available to the public through US-EPA website available at: http://www.epa.gov/ceampubl/swater/efdc/index.html. The EFDC can be used to simulate aquatic systems in multiple dimensions with the stretched or sigma vertical coordinates and the Cartesian (or curvilinear), and orthogonal horizontal coordinates to represent the physi‐ cal characteristics of a water body. A dynamically-coupled transport process for turbulent kinetic energy, turbulent length scale, salinity and temperature are included in the EFDC model. The EFDC allows for drying and wetting in shallow water bodies by a mass conser‐ vation scheme.

Refereed EFDC-model-based evaluations exist predominately for stream ecosystems. Exam‐ ples include a three-dimensional hydrodynamic model of the Chicago River, Illinois (Sinha et al., 2012); the effect of interacting downstream branches on saltwater intrusion in the Modaomen Estuary, China (Gong et al., 2012); and comparison of two hydrodynamic mod‐ els of Weeks Bay, Alabama (Alarcon et al., 2012).

ment. Currently, land-use data layers are available in geographic information systems (GIS)

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

63

The SWAT model also requires distributed detail soils data, which is available from either State Soil Geographic (STATSGO) database or Soil Survey Geographic (SSURGO) databases (USDA, 2005). The SSURGO database is the most detailed data source currently available in the U.S. as it provides more soil polygons per unit area. The DEM, soils, and landuse geo‐ graphic data layers should be all projected in one projection system (e.g. Universal Trans‐

Most of the watershed or field scale models (e.g. SWAT, WEPP) have embedded weather stations and climate generators. However, more field-specific climate inputs (e.g. rainfall; daily minimum, maximum and mean temperatures; solar radiation; relative humidity, and wind speed) can be allowed in the model for the watershed assessment. Weather data such as daily rainfall and ambient temperature can be downloaded from the National Climatic Data Center (NCDC, 2012). Other field-specific model input parameters such as irrigation (e.g. auto or manual irrigation), fertilizer application (application rates, fertilizer type), crop rotation (e.g. corn after soybean), tillage (e.g. conventional, reduced, no-tillage), planting

The major procedures in water quality modeling with HSPF are the construction of a conceptual model, mathematical description of the conceptual model, preparation of in‐ put data such as time series parameter values, calibration and validation of the model, and application of the model for field conditions. Time series input data can be sup‐ plied into the HSPF model by using a stand-alone program or the Watershed Data Management program (WDM) provided in BASINS (Better assessment science integrat‐ ing point and nonpoint sources). BASINS is a multipurpose environmental analysis sys‐ tem model, which can be utilized by regional, state, and local agencies for conducting water quality based studies. The BASINS system incorporates an open source geograph‐ ic information system (GIS) (i.e., MapWindow), the national watershed and meteorologi‐ cal data, and the state-of-the-art environmental models such as HSPF, Pollutant Loading

Application (PLOAD), and SWAT into one convenient package (USEPA, 2010).

Normally, the development of a HSPF model starts with a watershed delineation process, which includes the setup of digital elevation model (DEM) data in the ArcInfo grid format, generation of stream networks in shape format, and designation of watershed inlets or out‐ lets using the watershed delineation tool built in the BASINS. The HSPF also needs land use and soil data to determine the area and the hydrologic parameters of each land use pattern in the model, which can be done with the land use and soil classification tool in the BASINS. The HSPF is a lumped parameter model with a modular structure. The PERLAND modular represents the pervious land segments over which a considerable amount of water infiltrates into the ground. The IMPLND modular denotes the Impervious land segments over which infiltration is negligible such as paved urban surfaces. Processes involving water bodies like streams and lakes are represented with the RCHRES module. These modules have many

format, which is applicable for the watershed modeling.

and harvesting dates can be defined (Parajuli, 2010b).

verse Mercator-UTM 1983, zone 16).

**3.2. HSPF model**

## **2.8. SWMM**

The US-EPA's Storm Water Management Model (SWMM) was initially developed in 1971, and has been significantly upgraded (http://www.epa.gov/nrmrl/wswrd/wq/models / swmm/index.htm). The SWMM model is a widely used model for planning, analysis and design related to storm water runoff, sewers, and other drainage systems in urban areas. SWMM can simulate single storm-events or provide continuous prediction of surface-runoff quantity and quality from urban areas. In addition to predicting surface-runoff quantity and quality, the model can also predict flow rate, flow depth, and water quality in each pipe and channel.

There have been numerous applications of SWMM in the literature recently. Blumensaat et al. (2012) investigated sewer transport with SWMM under minimum data requirements. Cantone and Schmidt (2011) applied SWMM to improve understanding of the hydrologic response of highly urbanized watershed catchments like the Illinois Urban areas. Talei and Chua (2012) estimated the influence of lag-time on storm event-based hydrologic impacts (e.g. rainfall, surface-runoff) using the SWMM model and a data-driven approach.

## **3. Methods to develop a model**

Appropriate methods are needed to develop a model, utilize different data sources (e.g. dig‐ ital elevation, soil, land use, weather etc.), and develop methods to quantify pollutants source loads in the model. As examples, the methods development process is described here for two commonly used models (i.e., SWAT and HSPF).

## **3.1. SWAT model**

The SWAT model utilizes digital elevation model (DEM), soils, land cover, and weather da‐ ta such as precipitation, temperature, wind speed, solar radiation, and relative humidity. SWAT delineates watershed boundary and topographic characteristics of the watershed us‐ ing National Elevation Dataset called digital elevation model (DEM) data, which are availa‐ ble in the grid form with different resolutions (e.g. 30m x 30m grid; 10m x 10m grid) generally collected by U.S. Geological survey (USGS, 1999) or other sources. The 30m grid data are commonly used in the large scale watershed modeling work. However, small wa‐ tershed or field scale modeling may benefit from using of 10m x 10m resolution DEM data. Model defines land use inputs in the model are described using distributed land cover data (USDA-NASS, 2010) or other land use data. The time-specific land-cover data (e.g. 1992, 2001 and 2006) for the U.S. and Puerto Rico can be downloaded from the National Land Cover Database (NLCD), a publicly available data source. The distributed land cover data with land use classifications can provide essential model input for the watershed assess‐ ment. Currently, land-use data layers are available in geographic information systems (GIS) format, which is applicable for the watershed modeling.

The SWAT model also requires distributed detail soils data, which is available from either State Soil Geographic (STATSGO) database or Soil Survey Geographic (SSURGO) databases (USDA, 2005). The SSURGO database is the most detailed data source currently available in the U.S. as it provides more soil polygons per unit area. The DEM, soils, and landuse geo‐ graphic data layers should be all projected in one projection system (e.g. Universal Trans‐ verse Mercator-UTM 1983, zone 16).

Most of the watershed or field scale models (e.g. SWAT, WEPP) have embedded weather stations and climate generators. However, more field-specific climate inputs (e.g. rainfall; daily minimum, maximum and mean temperatures; solar radiation; relative humidity, and wind speed) can be allowed in the model for the watershed assessment. Weather data such as daily rainfall and ambient temperature can be downloaded from the National Climatic Data Center (NCDC, 2012). Other field-specific model input parameters such as irrigation (e.g. auto or manual irrigation), fertilizer application (application rates, fertilizer type), crop rotation (e.g. corn after soybean), tillage (e.g. conventional, reduced, no-tillage), planting and harvesting dates can be defined (Parajuli, 2010b).

## **3.2. HSPF model**

et al., 2012); the effect of interacting downstream branches on saltwater intrusion in the Modaomen Estuary, China (Gong et al., 2012); and comparison of two hydrodynamic mod‐

The US-EPA's Storm Water Management Model (SWMM) was initially developed in 1971, and has been significantly upgraded (http://www.epa.gov/nrmrl/wswrd/wq/models / swmm/index.htm). The SWMM model is a widely used model for planning, analysis and design related to storm water runoff, sewers, and other drainage systems in urban areas. SWMM can simulate single storm-events or provide continuous prediction of surface-runoff quantity and quality from urban areas. In addition to predicting surface-runoff quantity and quality, the model can also predict flow rate, flow depth, and water quality in each pipe and

There have been numerous applications of SWMM in the literature recently. Blumensaat et al. (2012) investigated sewer transport with SWMM under minimum data requirements. Cantone and Schmidt (2011) applied SWMM to improve understanding of the hydrologic response of highly urbanized watershed catchments like the Illinois Urban areas. Talei and Chua (2012) estimated the influence of lag-time on storm event-based hydrologic impacts

Appropriate methods are needed to develop a model, utilize different data sources (e.g. dig‐ ital elevation, soil, land use, weather etc.), and develop methods to quantify pollutants source loads in the model. As examples, the methods development process is described here

The SWAT model utilizes digital elevation model (DEM), soils, land cover, and weather da‐ ta such as precipitation, temperature, wind speed, solar radiation, and relative humidity. SWAT delineates watershed boundary and topographic characteristics of the watershed us‐ ing National Elevation Dataset called digital elevation model (DEM) data, which are availa‐ ble in the grid form with different resolutions (e.g. 30m x 30m grid; 10m x 10m grid) generally collected by U.S. Geological survey (USGS, 1999) or other sources. The 30m grid data are commonly used in the large scale watershed modeling work. However, small wa‐ tershed or field scale modeling may benefit from using of 10m x 10m resolution DEM data. Model defines land use inputs in the model are described using distributed land cover data (USDA-NASS, 2010) or other land use data. The time-specific land-cover data (e.g. 1992, 2001 and 2006) for the U.S. and Puerto Rico can be downloaded from the National Land Cover Database (NLCD), a publicly available data source. The distributed land cover data with land use classifications can provide essential model input for the watershed assess‐

(e.g. rainfall, surface-runoff) using the SWMM model and a data-driven approach.

els of Weeks Bay, Alabama (Alarcon et al., 2012).

62 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

**3. Methods to develop a model**

**3.1. SWAT model**

for two commonly used models (i.e., SWAT and HSPF).

**2.8. SWMM**

channel.

The major procedures in water quality modeling with HSPF are the construction of a conceptual model, mathematical description of the conceptual model, preparation of in‐ put data such as time series parameter values, calibration and validation of the model, and application of the model for field conditions. Time series input data can be sup‐ plied into the HSPF model by using a stand-alone program or the Watershed Data Management program (WDM) provided in BASINS (Better assessment science integrat‐ ing point and nonpoint sources). BASINS is a multipurpose environmental analysis sys‐ tem model, which can be utilized by regional, state, and local agencies for conducting water quality based studies. The BASINS system incorporates an open source geograph‐ ic information system (GIS) (i.e., MapWindow), the national watershed and meteorologi‐ cal data, and the state-of-the-art environmental models such as HSPF, Pollutant Loading Application (PLOAD), and SWAT into one convenient package (USEPA, 2010).

Normally, the development of a HSPF model starts with a watershed delineation process, which includes the setup of digital elevation model (DEM) data in the ArcInfo grid format, generation of stream networks in shape format, and designation of watershed inlets or out‐ lets using the watershed delineation tool built in the BASINS. The HSPF also needs land use and soil data to determine the area and the hydrologic parameters of each land use pattern in the model, which can be done with the land use and soil classification tool in the BASINS. The HSPF is a lumped parameter model with a modular structure. The PERLAND modular represents the pervious land segments over which a considerable amount of water infiltrates into the ground. The IMPLND modular denotes the Impervious land segments over which infiltration is negligible such as paved urban surfaces. Processes involving water bodies like streams and lakes are represented with the RCHRES module. These modules have many components dealing with hydrological and water quality processes. Detailed information about the structure and functioning of these modules can be found in elsewhere (Donigian and Crawford 1976; Donigian et al. 1984; Bicknel et al. 1993; Chen et al. 1998).

## **4. Model application**

Two watersheds in Mississippi (Upper Pearl River and Yazoo River Basin) were selected for modeling case studies using two hydrologic and water quality models (SWAT and HSPF). Models were calibrated and validated using USGS observed streamflow data for the current conditions and models were applied to predict future climate change scenarios impact on hydrology. Case studies demonstrated how future climate change scenarios impact stream‐ flow from the watersheds.

## **4.1. SWAT model**

The main objective of this case study was to quantify the potential impact of future climate change scenarios on hydrologic characteristics such as monthly average streamflow with in the Upper Pearl River Watershed (UPRW) using the SWAT model. The specific objectives were to: (1) develop a site-specific SWAT model for the UPRW based on watershed charac‐ teristics, climatic, and hydrological conditions; (2) calibrate and validate model using USGS observed stream flow data; and (3) develop future climate change scenarios and quantify their impacts on stream flows.

**Figure 1.** Location map of the watershed showing sub-watersheds and others

The SWAT model predicted monthly streamflow values were compared separately for model calibration and validation periods using three common parameters (coefficient of

RMSE). The monthly model performances were ranked excellent for R2 or E values > 0.90, very good for values between 0.75–0.89, good for values between 0.50–0.74, fair for values between 0.25–0.49, poor for values between 0–0.24, and unsatisfactory for values < 0 (Moriasi et al., 2007; Parajuli et al., 2008, 2009). The RMSE performance has no suggest‐ ed values to rank, however the smaller the RMSE the better the performance of the mod‐ el (Moriasi et al., 2007), and a value of zero for RMSE represents perfect simulation of

The SWAT model was calibrated (from January 1998 to December 2003) and validated (from January 2004 to December 2009) using field observed monthly streamflow data from the Lena USGS gage station (USGS 02483500) within the UPRW. Model calibration and validation parameters were adopted from previous study (Parajuli, 2010a). Model si‐ mulated results showed good to very good performances for the monthly streamflow prediction both during model calibration (R2 = 0.75, E = 0.70) and validation (R2 = 0.73, E = 0.51) periods (Fig. 2). The SWAT model predicted monthly streamflow (m3 s-1) estimat‐ ed very similar RMSE values (<2% difference) during model calibration (RMSE = 51.7 m3 s-1) and validation (RMSE = 50.7 m3 s-1) periods. This case study results were in close agreement with several previous studies that used the SWAT model (Gassman et al., 2007; Moriasi et al., 2007; Parajuli et al., 2009; Parajuli 2010a; Nejadhashemi et al., 2011;

; Nash–Sutcliffe efficiency index – E; and root mean square error -

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

65

*4.1.2. Model calibration and validation*

determination – R2

the measured data.

Sheshukov et al., 2011).

#### *4.1.1. Study area and model development*

The SWAT model was developed and applied in the UPRW (7,588 km2 ), which is located in Mississippi (Fig. 1). The UPRW covers ten counties (Choctaw, Attala, Winston, Leake, Ne‐ sobha, Kemper, Madison, Rankin, Scott and Newton) in Mississippi with predominant land uses of woodland (72%), grassland (20%), urban land (6%) and others (2%).

To develop the SWAT model, this case study utilized national elevation data, which is also called DEM data of 30m x 30m grids to delineate watershed boundary. The STATSGO was used to create distributed soil data input in the model. The land cover data was created us‐ ing the cropland data layer in the model. The climate data (e.g. daily precipitation, tempera‐ ture) were used from several weather stations within or near the watershed as maintained by the National Climatic Data Center. The SWAT model allows several potential evapo‐ transpiration estimation method alternatives (e.g. Penman-Monteith, Hargreaves, Priestley-Taylor). This case study utilized the Penman-Monteith method to estimate PET, which requires daily rainfall, maximum and minimum temperatures, relative-humidity, solar radi‐ ation, and wind speed data. The additional data needed to simulate the SWAT model using Penman-Monteith PET method were generated by the SWAT model.

**Figure 1.** Location map of the watershed showing sub-watersheds and others

#### *4.1.2. Model calibration and validation*

components dealing with hydrological and water quality processes. Detailed information about the structure and functioning of these modules can be found in elsewhere (Donigian

Two watersheds in Mississippi (Upper Pearl River and Yazoo River Basin) were selected for modeling case studies using two hydrologic and water quality models (SWAT and HSPF). Models were calibrated and validated using USGS observed streamflow data for the current conditions and models were applied to predict future climate change scenarios impact on hydrology. Case studies demonstrated how future climate change scenarios impact stream‐

The main objective of this case study was to quantify the potential impact of future climate change scenarios on hydrologic characteristics such as monthly average streamflow with in the Upper Pearl River Watershed (UPRW) using the SWAT model. The specific objectives were to: (1) develop a site-specific SWAT model for the UPRW based on watershed charac‐ teristics, climatic, and hydrological conditions; (2) calibrate and validate model using USGS observed stream flow data; and (3) develop future climate change scenarios and quantify

Mississippi (Fig. 1). The UPRW covers ten counties (Choctaw, Attala, Winston, Leake, Ne‐ sobha, Kemper, Madison, Rankin, Scott and Newton) in Mississippi with predominant land

To develop the SWAT model, this case study utilized national elevation data, which is also called DEM data of 30m x 30m grids to delineate watershed boundary. The STATSGO was used to create distributed soil data input in the model. The land cover data was created us‐ ing the cropland data layer in the model. The climate data (e.g. daily precipitation, tempera‐ ture) were used from several weather stations within or near the watershed as maintained by the National Climatic Data Center. The SWAT model allows several potential evapo‐ transpiration estimation method alternatives (e.g. Penman-Monteith, Hargreaves, Priestley-Taylor). This case study utilized the Penman-Monteith method to estimate PET, which requires daily rainfall, maximum and minimum temperatures, relative-humidity, solar radi‐ ation, and wind speed data. The additional data needed to simulate the SWAT model using

), which is located in

The SWAT model was developed and applied in the UPRW (7,588 km2

uses of woodland (72%), grassland (20%), urban land (6%) and others (2%).

Penman-Monteith PET method were generated by the SWAT model.

and Crawford 1976; Donigian et al. 1984; Bicknel et al. 1993; Chen et al. 1998).

64 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

**4. Model application**

flow from the watersheds.

their impacts on stream flows.

*4.1.1. Study area and model development*

**4.1. SWAT model**

The SWAT model predicted monthly streamflow values were compared separately for model calibration and validation periods using three common parameters (coefficient of determination – R2 ; Nash–Sutcliffe efficiency index – E; and root mean square error - RMSE). The monthly model performances were ranked excellent for R2 or E values > 0.90, very good for values between 0.75–0.89, good for values between 0.50–0.74, fair for values between 0.25–0.49, poor for values between 0–0.24, and unsatisfactory for values < 0 (Moriasi et al., 2007; Parajuli et al., 2008, 2009). The RMSE performance has no suggest‐ ed values to rank, however the smaller the RMSE the better the performance of the mod‐ el (Moriasi et al., 2007), and a value of zero for RMSE represents perfect simulation of the measured data.

The SWAT model was calibrated (from January 1998 to December 2003) and validated (from January 2004 to December 2009) using field observed monthly streamflow data from the Lena USGS gage station (USGS 02483500) within the UPRW. Model calibration and validation parameters were adopted from previous study (Parajuli, 2010a). Model si‐ mulated results showed good to very good performances for the monthly streamflow prediction both during model calibration (R2 = 0.75, E = 0.70) and validation (R2 = 0.73, E = 0.51) periods (Fig. 2). The SWAT model predicted monthly streamflow (m3 s-1) estimat‐ ed very similar RMSE values (<2% difference) during model calibration (RMSE = 51.7 m3 s-1) and validation (RMSE = 50.7 m3 s-1) periods. This case study results were in close agreement with several previous studies that used the SWAT model (Gassman et al., 2007; Moriasi et al., 2007; Parajuli et al., 2009; Parajuli 2010a; Nejadhashemi et al., 2011; Sheshukov et al., 2011).

tation (-20%). However, Sc14 had the greatest effect on stream flow among all low condition scenarios, due to the highest temperature (+2 degree Celsius) and CO2 values (660 ppmv).

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

67

**Table 1.** Simulated climate change parameters scenarios and effect

**Predicted cumulative stream flow (m3s-1**

**)**

**1**

**9**

**17**

**25**

**33**

**41**

**49**

**57**

**65**

than base condition and lower than base condition scenarios.

**73**

**81**

**89**

**97**

**105**

**113**

**121**

**129**

**137**

**145**

**153**

**161**

**169**

**Figure 3.** Model predicted cumulative monthly streamflow during thirty years period (2010-2040) showing greater

**177**

**Months (1 to 372) from Januray 2010 to December 2040** 

**185**

**193**

**201**

**209**

**217**

**225**

**233**

**241**

**249**

**257**

**265**

**273**

**281**

**289**

**297**

**305**

**313**

**321**

Sc 13

**329**

**337**

Sc14

**345**

**353**

Sc12

Base

**361**

**369**

**Figure 2.** Monthly observed vs. predicted streamflow during (a) calibration and (b) validation periods

#### *4.1.3. Future climate scenarios*

The calibrated and validated SWAT model for the UPRW was simulated for an additional 30 years (January 2010 to December 2040) to provide fourteen future climate change sce‐ narios (Table 1). The average streamflow value from the calibrated and validated model was considered as baseline scenario. The future climate change scenarios represented per‐ centage change in the precipitation, temperature and CO2 concentration values as descri‐ bed in the Table 1. The CO2 values were adjusted from a baseline value of 330 ppmv (part per million by volume), which is a default value provided in the SWAT model. Two other CO2 values (495 and 660) were tested in the model considering 50% and 100% increase from the model default value. Percentage changes in the precipitation were simulated for ±20% from the baseline value. Similarly, the model temperature factor was adjusted using +1 and +2 degrees in Celsius from the baseline. The fourteen future climate change scenar‐ ios were developed using interaction of three CO2, three precipitation, and three tempera‐ ture adjustment values.

The SWAT model results for fourteen scenarios (from Sc1 to Sc14 for Lena gage station) predicted an average maximum monthly stream flow decrease of 57% and average maxi‐ mum monthly flow increase of 74% from the base simulation (Figure 3). Precipitation in‐ crease always had the greatest impact on monthly streamflow from the watershed. A twenty percent increase in precipitation resulted into the greatest impact in the future streamflow prediction. However, increases in CO2 and temperature accelerated the magni‐ tude of streamflow process.

Scenario 13 with the highest increase in the precipitation (+20%), CO2 (660 ppmv), and tem‐ perature (+2 degree Celsius) had about 74% greater impact on streamflow prediction than the baseline condition (Fig. 3). Other scenarios that had high impact on streamflow predic‐ tion were Sc1, Sc4, Sc7, and Sc10. The increase in the temperature had medium impact on streamflow process as shown by Sc3, Sc6, Sc9, and Sc12. However, Sc12 had the greatest im‐ pact among medium scenarios as it predicted about 10% greater cumulative monthly streamflow than the baseline condition. Scenarios Sc2, Sc5, Sc8, Sc11, and Sc14 had lower cu‐ mulative monthly streamflow than the baseline condition, as they all had decreased precipi‐


tation (-20%). However, Sc14 had the greatest effect on stream flow among all low condition scenarios, due to the highest temperature (+2 degree Celsius) and CO2 values (660 ppmv).

**Table 1.** Simulated climate change parameters scenarios and effect

**y = 0.82x + 42.10 R² = 0.75**

66 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

(a) **y = 0.95x + 34.92**

(b)

**Predicted monthly flow (m3**

The calibrated and validated SWAT model for the UPRW was simulated for an additional 30 years (January 2010 to December 2040) to provide fourteen future climate change sce‐ narios (Table 1). The average streamflow value from the calibrated and validated model was considered as baseline scenario. The future climate change scenarios represented per‐ centage change in the precipitation, temperature and CO2 concentration values as descri‐ bed in the Table 1. The CO2 values were adjusted from a baseline value of 330 ppmv (part per million by volume), which is a default value provided in the SWAT model. Two other CO2 values (495 and 660) were tested in the model considering 50% and 100% increase from the model default value. Percentage changes in the precipitation were simulated for ±20% from the baseline value. Similarly, the model temperature factor was adjusted using +1 and +2 degrees in Celsius from the baseline. The fourteen future climate change scenar‐ ios were developed using interaction of three CO2, three precipitation, and three tempera‐

The SWAT model results for fourteen scenarios (from Sc1 to Sc14 for Lena gage station) predicted an average maximum monthly stream flow decrease of 57% and average maxi‐ mum monthly flow increase of 74% from the base simulation (Figure 3). Precipitation in‐ crease always had the greatest impact on monthly streamflow from the watershed. A twenty percent increase in precipitation resulted into the greatest impact in the future streamflow prediction. However, increases in CO2 and temperature accelerated the magni‐

Scenario 13 with the highest increase in the precipitation (+20%), CO2 (660 ppmv), and tem‐ perature (+2 degree Celsius) had about 74% greater impact on streamflow prediction than the baseline condition (Fig. 3). Other scenarios that had high impact on streamflow predic‐ tion were Sc1, Sc4, Sc7, and Sc10. The increase in the temperature had medium impact on streamflow process as shown by Sc3, Sc6, Sc9, and Sc12. However, Sc12 had the greatest im‐ pact among medium scenarios as it predicted about 10% greater cumulative monthly streamflow than the baseline condition. Scenarios Sc2, Sc5, Sc8, Sc11, and Sc14 had lower cu‐ mulative monthly streamflow than the baseline condition, as they all had decreased precipi‐

**Figure 2.** Monthly observed vs. predicted streamflow during (a) calibration and (b) validation periods

 **s-1**

**)**

**R² = 0.73**

**E = 0.51**

**0 40 80 120 160 200 240 280 320 360 400**

**Observed monthly flow (m3 s-1)**

**E = 0.70**

**0 40 80 120 160 200 240 280 320 360 400**

**Observed monthly flow (m3 s-1)**

*4.1.3. Future climate scenarios*

ture adjustment values.

tude of streamflow process.

**Predicted monthly flow (m3**

 **s-1**

**)**

**Figure 3.** Model predicted cumulative monthly streamflow during thirty years period (2010-2040) showing greater than base condition and lower than base condition scenarios.

## **4.2. HSPF model**

The goal of this case study was to estimate the potential impact of future climate change upon hydrologic characteristics such as river discharge, surface evaporation, and water out‐ flow in the YRB (Yazoo River Basin) using the HSPF model. The specific objectives were to: (1) develop a site-specific model for the YRB based on watershed, meteorological, and hy‐ drological conditions; (2) calibrate the resulting model using existing field data and/or com‐ putational data; and (3) create simulation scenarios to project the potential impact of future climate changes upon hydrologic characteristics in the YRB.

of air temperature from high to low order as: MIROC32A1B > CSIROMK35A1B > CSIR‐

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

69

The HSPF model for this case study was developed using the PERLND, IMPLND, and RCHRES modules that are available in HSPF. The PWATER section of the PERLND module is a major component that simulates the water budget, including surface flow, inter-flow and groundwater behavior. The HYDR section of the RCHRES module simulates the hy‐

**Mississippi**

**N**

OMK32A1B > HADCM3B2.

draulic behavior of the stream.

**Basin Outlet**

**Figure 4.** Location of modeled area in the Yazoo River Basin, Mississippi.

#### *4.2.1. Study area and model development*

The YRB is the largest river basin in Mississippi, USA and has a total drainage area of 34,600 km2 (Fig. 4). This basin is separated into two distinct topographic regions, one is the Bluff Hills (about 16600 km2 ) and the other is the Mississippi Alluvial Delta (Guedon and Tho‐ mas, 2004; MDEQ, 2008; Shields et al., 2008). The Bluff Hills region is a hilly and upland area where streams originate from lush oak and hickory forests and pastures dominate the rural landscape. The Delta Region, on the other hand, is a flat and lowland area characterized by slow streamflow and an extensive system of oxbow lakes.

Data collection for the YRB (HUC 8030208) includes watershed descriptions, meteorological, and hydrologic data. Several agencies are active in the data collection efforts. Most of the data used in this study such as land use, soil type, topography, precipitation, and discharge are from National Hydrography Dataset, U.S. Geologic Survey National Water Information System, and 2001 National Land Cover Data.

Four future climate change scenario data, namely the HADCM3B2, CSIROMK35A1B, CSIR‐ OMK2A1B, and MIROC32A1B, were used in this case study. HADCM3, CSIROMK35, CSIR‐ OMK2, and MIROC32 are names of climate general circulation models (GCM). The B2 and A1B at the end of the names of the climate change scenarios are the Intergovernmental Panel on Climate Change (IPCC) emission scenarios under which the GCMs were run to produce the individual climate projection. The HADCM3B2 scenario data was obtained from the Hadley Centre for Climate Prediction and Research, United Kingdom. The CSIROMK35A1B and CSIROMK2A1B scenarios data were obtained from the Australian Commonwealth Sci‐ entific and Research Organization Atmospheric Research, and the MIROC32A1B scenario data was obtained from the Center for Climate System Research, University of Tokyo Na‐ tional Institute for Environmental Studies and Frontier Research Center for Global Change. More detail information about these climate scenarios are available at: http://wwwpcmdi.llnl.gov/ipcc/model\_documentation/ipcc\_model\_documentation.php. These four sce‐ narios data involve monthly air temperature and precipitation for a period from 2000 to 2050, which were generated by GCMs and the Center for Climate System Research National Institute for Environmental Studies and Frontier Research Center for Global Change (Uni‐ versity of Tokyo). These data were scaled to the 8-digit HUC watersheds for different re‐ gions. For the YRB watershed, the 8-digit HUC was 08030208. A descriptive statistics for these four scenarios data showed the amount of precipitation from high to low order as: CSIROMK35A1B > HADCM3B2 > CSIROMK2A1B > MIROC32A1B, whereas the magnitude of air temperature from high to low order as: MIROC32A1B > CSIROMK35A1B > CSIR‐ OMK32A1B > HADCM3B2.

**4.2. HSPF model**

The goal of this case study was to estimate the potential impact of future climate change upon hydrologic characteristics such as river discharge, surface evaporation, and water out‐ flow in the YRB (Yazoo River Basin) using the HSPF model. The specific objectives were to: (1) develop a site-specific model for the YRB based on watershed, meteorological, and hy‐ drological conditions; (2) calibrate the resulting model using existing field data and/or com‐ putational data; and (3) create simulation scenarios to project the potential impact of future

The YRB is the largest river basin in Mississippi, USA and has a total drainage area of 34,600 km2 (Fig. 4). This basin is separated into two distinct topographic regions, one is the Bluff

mas, 2004; MDEQ, 2008; Shields et al., 2008). The Bluff Hills region is a hilly and upland area where streams originate from lush oak and hickory forests and pastures dominate the rural landscape. The Delta Region, on the other hand, is a flat and lowland area characterized by

Data collection for the YRB (HUC 8030208) includes watershed descriptions, meteorological, and hydrologic data. Several agencies are active in the data collection efforts. Most of the data used in this study such as land use, soil type, topography, precipitation, and discharge are from National Hydrography Dataset, U.S. Geologic Survey National Water Information

Four future climate change scenario data, namely the HADCM3B2, CSIROMK35A1B, CSIR‐ OMK2A1B, and MIROC32A1B, were used in this case study. HADCM3, CSIROMK35, CSIR‐ OMK2, and MIROC32 are names of climate general circulation models (GCM). The B2 and A1B at the end of the names of the climate change scenarios are the Intergovernmental Panel on Climate Change (IPCC) emission scenarios under which the GCMs were run to produce the individual climate projection. The HADCM3B2 scenario data was obtained from the Hadley Centre for Climate Prediction and Research, United Kingdom. The CSIROMK35A1B and CSIROMK2A1B scenarios data were obtained from the Australian Commonwealth Sci‐ entific and Research Organization Atmospheric Research, and the MIROC32A1B scenario data was obtained from the Center for Climate System Research, University of Tokyo Na‐ tional Institute for Environmental Studies and Frontier Research Center for Global Change. More detail information about these climate scenarios are available at: http://wwwpcmdi.llnl.gov/ipcc/model\_documentation/ipcc\_model\_documentation.php. These four sce‐ narios data involve monthly air temperature and precipitation for a period from 2000 to 2050, which were generated by GCMs and the Center for Climate System Research National Institute for Environmental Studies and Frontier Research Center for Global Change (Uni‐ versity of Tokyo). These data were scaled to the 8-digit HUC watersheds for different re‐ gions. For the YRB watershed, the 8-digit HUC was 08030208. A descriptive statistics for these four scenarios data showed the amount of precipitation from high to low order as: CSIROMK35A1B > HADCM3B2 > CSIROMK2A1B > MIROC32A1B, whereas the magnitude

) and the other is the Mississippi Alluvial Delta (Guedon and Tho‐

climate changes upon hydrologic characteristics in the YRB.

68 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

slow streamflow and an extensive system of oxbow lakes.

System, and 2001 National Land Cover Data.

*4.2.1. Study area and model development*

Hills (about 16600 km2

The HSPF model for this case study was developed using the PERLND, IMPLND, and RCHRES modules that are available in HSPF. The PWATER section of the PERLND module is a major component that simulates the water budget, including surface flow, inter-flow and groundwater behavior. The HYDR section of the RCHRES module simulates the hy‐ draulic behavior of the stream.

**Figure 4.** Location of modeled area in the Yazoo River Basin, Mississippi.

## *4.2.2. Model calibration and validation*

Model calibration involves adjusting input parameters within a reasonable range to obtain a best fitness between field observations and model predictions. Model validation is a process of validating the calibrated model by comparing the field observations against the model predictions without changing any input parameter values. Table 2 shows a comparison of the observed and predicted annual water outflow volume. The annual differences in errors between the observed and predicted water outflow volumes were about 6% and were, there‐ fore, acceptable (Bicknell et al., 2001). With prediction = 0.97\*observation and R2 = 0.98 and E = 0.96, we determined that an excellent agreement was obtained between the field obser‐ vations and model predictions during the model calibration process.

*4.2.3. Past and future climate change*

evaporative losses will respond in the future.

simulation period (Table 3).

and future 10 years.

Comparison of mean annual water yields between the past 10 years (2001-2011) and the fu‐ ture 40 years (2011-2050) for the four climate projections indicates that water yields will con‐ tinue to decline (Table 3). The percent change in mean annual water yield varied from 29.47% for the CSIROMK35A1B projection to 18.51% for the MIROC32A1B projection, with four climate projections indicating continuing declines out to 2050. The same decline trends were observed for maximum annual water yields (Table 3). The declines in mean and maxi‐ mum annual water yields occurred primarily due to the projected precipitation decrease. Mixed results were found for the mean annual evaporative loss (Table 3). The CSIROMK2B2 projection indicated a long-term increase while the other three projections indicated a longterm decrease in evaporative losses. Further work is thus necessary to better determine how

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

71

Changes of monthly minimum, mean, and maximum in water discharges and yields for the four climate projections during the 40-year simulation period (2011-2050) are given in Figs. 5 and 6. The monthly minimum, mean, and maximum water discharges and yields varied among the four climate projections and changed from year to year within each projection. In general, the MIROC32A1B projection had highest monthly minimum, mean, and maximum water discharges and yields in most of the years during the 40-year simulation, which occur‐ red because the MIROC32A1B projection had highest annual precipitation during the same

**Table 3.** Comparison of the sum and mean values for precipitation, evaporative loss, and water yield between the past

Comparison of annual water outflow between the observations and predictions for a time period from January 1, 2005 to December 31, 2010 during the model validation process was given in Table 2. The regression equation predictions = 0.97\*observation and R2 = 0.99 and E = 0.97 verified the excellent agreement between the model predictions and the field observa‐ tions during the model validation process.


**Table 2.** Comparison of the simulated and observed annual water outflow volumes during model calibration and validation.

## *4.2.3. Past and future climate change*

*4.2.2. Model calibration and validation*

tions during the model validation process.

validation.

Model calibration involves adjusting input parameters within a reasonable range to obtain a best fitness between field observations and model predictions. Model validation is a process of validating the calibrated model by comparing the field observations against the model predictions without changing any input parameter values. Table 2 shows a comparison of the observed and predicted annual water outflow volume. The annual differences in errors between the observed and predicted water outflow volumes were about 6% and were, there‐ fore, acceptable (Bicknell et al., 2001). With prediction = 0.97\*observation and R2 = 0.98 and E = 0.96, we determined that an excellent agreement was obtained between the field obser‐

Comparison of annual water outflow between the observations and predictions for a time period from January 1, 2005 to December 31, 2010 during the model validation process was given in Table 2. The regression equation predictions = 0.97\*observation and R2 = 0.99 and E = 0.97 verified the excellent agreement between the model predictions and the field observa‐

**Table 2.** Comparison of the simulated and observed annual water outflow volumes during model calibration and

vations and model predictions during the model calibration process.

70 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

Comparison of mean annual water yields between the past 10 years (2001-2011) and the fu‐ ture 40 years (2011-2050) for the four climate projections indicates that water yields will con‐ tinue to decline (Table 3). The percent change in mean annual water yield varied from 29.47% for the CSIROMK35A1B projection to 18.51% for the MIROC32A1B projection, with four climate projections indicating continuing declines out to 2050. The same decline trends were observed for maximum annual water yields (Table 3). The declines in mean and maxi‐ mum annual water yields occurred primarily due to the projected precipitation decrease. Mixed results were found for the mean annual evaporative loss (Table 3). The CSIROMK2B2 projection indicated a long-term increase while the other three projections indicated a longterm decrease in evaporative losses. Further work is thus necessary to better determine how evaporative losses will respond in the future.

Changes of monthly minimum, mean, and maximum in water discharges and yields for the four climate projections during the 40-year simulation period (2011-2050) are given in Figs. 5 and 6. The monthly minimum, mean, and maximum water discharges and yields varied among the four climate projections and changed from year to year within each projection. In general, the MIROC32A1B projection had highest monthly minimum, mean, and maximum water discharges and yields in most of the years during the 40-year simulation, which occur‐ red because the MIROC32A1B projection had highest annual precipitation during the same simulation period (Table 3).


**Table 3.** Comparison of the sum and mean values for precipitation, evaporative loss, and water yield between the past and future 10 years.

**5. Conclusions**

SWAT and HSPF models.

the YRB.

dated SWAT model provided good to very good fits (R2

year's model simulation period in this study.

impact on hydrology using two models.

Two models (SWAT and HSPF) commonly used in hydrological and water quality studies were applied here in two large scale watersheds (UPRW and YRB) in the state of Mississip‐ pi. Models were calibrated and validated using USGS observed streamflow data. The longterm hydrological impacts due to future climate change scenarios were assessed using the

For one case study, simulated mean monthly streamflow results for the calibrated and vali‐

USGS monthly observed streamflow data. Fourteen future climate change scenarios were developed using interaction of precipitation, CO2, and temperature adjustment values in the SWAT model. The scenario with the highest increase in the precipitation (+20%), CO2 (660 ppmv), and temperature (+2 degree Celsius) had about the greatest (> 74%) impact on streamflow simulation when compare with the baseline condition. Interaction of tempera‐ ture adjustment and CO2 factors had a medium and low impact respectively during thirty

Another case study examined the impact of climate change on future water discharge, evap‐ oration, and yield in the YRB using the BASINS-HSPF model. The model was calibrated us‐ ing observed data from a five-year (2001 to 2004), and validated using observed data from another five-year (2005 to 2010). Excellent agreements were obtained between the model

Four future climate scenarios (or projections) - CSIROMK35A1B, HADCM3B2, CSIR‐ OMK2B2, and MIROC32A1B were used to investigate water discharge, evaporative loss, and water outflow responses to predicted precipitation and air temperature changes over a 50-year period from 2001 to 2050. Comparison of simulation results between the past 10 years (2001-2010) and the future 40 years (2011-2050) shows that the mean and maximum annual water yields declined due to the projected precipitation decrease. In general, the MIROC32A1B projection had the highest monthly minimum, mean, and maximum water discharges and yields in most of the years during the 40-year simulation period (2011-2050). This projection had the highest projected annual precipitation. Results suggest that the projected precipitation had profound impacts upon water discharge and yield in

Spatial data used in the models may have potential sources of errors. For example, the DEM data are used to delineate watershed boundary are available in different resolutions. Simi‐ larly, use of land use, soils and weather data may have some spatial errors, which can influ‐ ence the hydrologic and climate change impact. However, these results will only have relative influence in model simulated results. This book chapter provided review of several watershed and water quality models and two case studies to evaluate future climate change

predictions and the field observations for model calibration and validation.

and E values from 0.75 to 0.51) to

http://dx.doi.org/10.5772/53596

73

Watershed-Scale Hydrological Modeling Methods and Applications

**Figure 5.** Simulated monthly minimum (a), mean (b), and maximum (c) discharge for the four simulation scenarios.

**Figure 6.** Simulated monthly minimum (a), mean (b), and maximum (c) water outflow volume for the four simulation scenarios.

## **5. Conclusions**

**2011 2020 2030 2040 2050**

**2011 2020 2030 2040 2050**

**Figure 6.** Simulated monthly minimum (a), mean (b), and maximum (c) water outflow volume for the four simulation

**Figure 5.** Simulated monthly minimum (a), mean (b), and maximum (c) discharge for the four simulation scenarios.

**CSIRO-A1B CSIRO-B2 (a)**

**HAD-B2 MIROC-A1B CSIRO-A1B CSIRO-B2**

**HAD-B2 MIROC-A1B**

**(a)**

72 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

**(b)**

**(c)**

**Monthly minimum (a), mean (b), and maximum (c) water yield (m3)**

scenarios.

**0.0E+00**

**2.0E+05**

**4.0E+05**

**(c)**

**(b)**

**6.0E+05**

**0.0E+00**

**2.0E+05**

**4.0E+05**

**6.0E+05**

**0.0E+00**

**2.0E+05**

**4.0E+05**

**6.0E+05**

**Monthly minimum (a), mean (b), and maximum (c) discharge (m3/s)**

Two models (SWAT and HSPF) commonly used in hydrological and water quality studies were applied here in two large scale watersheds (UPRW and YRB) in the state of Mississip‐ pi. Models were calibrated and validated using USGS observed streamflow data. The longterm hydrological impacts due to future climate change scenarios were assessed using the SWAT and HSPF models.

For one case study, simulated mean monthly streamflow results for the calibrated and vali‐ dated SWAT model provided good to very good fits (R2 and E values from 0.75 to 0.51) to USGS monthly observed streamflow data. Fourteen future climate change scenarios were developed using interaction of precipitation, CO2, and temperature adjustment values in the SWAT model. The scenario with the highest increase in the precipitation (+20%), CO2 (660 ppmv), and temperature (+2 degree Celsius) had about the greatest (> 74%) impact on streamflow simulation when compare with the baseline condition. Interaction of tempera‐ ture adjustment and CO2 factors had a medium and low impact respectively during thirty year's model simulation period in this study.

Another case study examined the impact of climate change on future water discharge, evap‐ oration, and yield in the YRB using the BASINS-HSPF model. The model was calibrated us‐ ing observed data from a five-year (2001 to 2004), and validated using observed data from another five-year (2005 to 2010). Excellent agreements were obtained between the model predictions and the field observations for model calibration and validation.

Four future climate scenarios (or projections) - CSIROMK35A1B, HADCM3B2, CSIR‐ OMK2B2, and MIROC32A1B were used to investigate water discharge, evaporative loss, and water outflow responses to predicted precipitation and air temperature changes over a 50-year period from 2001 to 2050. Comparison of simulation results between the past 10 years (2001-2010) and the future 40 years (2011-2050) shows that the mean and maximum annual water yields declined due to the projected precipitation decrease. In general, the MIROC32A1B projection had the highest monthly minimum, mean, and maximum water discharges and yields in most of the years during the 40-year simulation period (2011-2050). This projection had the highest projected annual precipitation. Results suggest that the projected precipitation had profound impacts upon water discharge and yield in the YRB.

Spatial data used in the models may have potential sources of errors. For example, the DEM data are used to delineate watershed boundary are available in different resolutions. Simi‐ larly, use of land use, soils and weather data may have some spatial errors, which can influ‐ ence the hydrologic and climate change impact. However, these results will only have relative influence in model simulated results. This book chapter provided review of several watershed and water quality models and two case studies to evaluate future climate change impact on hydrology using two models.

## **Acknowledgement**

The part of the case study research presented in this book chapter is based on work support‐ ed by the Special Research Initiatives (SRI) and Mississippi Agricultural and Forestry Ex‐ periment Station (MAFES) at Mississippi State University.

[7] Betrie G. D., Y. A. Mohamed A. van Griensven and R. Srinivasan. 2011. Sediment management modelling in the Blue Nile Basin using SWAT model. Hydrology and

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

75

[8] Bicknell B. R., Imhoff J. C., Kittle J. L., Donigian A. S., and Johanson R. C. 1993. Hy‐ drological Simulation Program – FORTRAN (HSPF): Users Manual for Release 10.

[9] Bicknell, B. R., Imhoff, J. C., Kittle, Jr., J. L., Jobes, T. H., Donigian, Jr., A. S., 2001. Hy‐ drological Simulation Program – Fortran, HSPF, Version 12, User's Manual. National Exposure Research Laboratory, Office of Research and Development, U.S. Environ‐

[10] Bingner R. L., and Theurer F. D. 2003. AnnAGNPS technical processes documenta‐ tion, Version 3.2. USDA-ARS, National Sedimentation Laboratory: Oxford, MS. [11] Blumensaat F., Wolfram M., and Krebs P. 2012. Sewer model development under minimum data requirements. Environmental Earth Sciences, 65:1427-1437.

[12] Boskidis, I., G. D. Gikas, G. K. Sylalos and V. A. Tsihrintzis. 2012. Water Resources

[13] Brown, J. B. 2011. Application of the SPARROW watershed model to describe nu‐ trient sources and transport in the Missouri River Basin: U.S. Geological Survey Fact

[14] Cantone, J., and Schmidt, A 2011. Improved understanding and prediction of the hy‐ drologic response of highly urbanized catchments through development of the Illi‐

[15] Chaubey, I., L. Chiang, M. W. Gitau, and S. Mohamed. 2010. Journal of soil and wa‐

[16] Chen Y. D., Carsel R. F., Mccutcheon S. C., and Nutter W. L. 1998. Stream tempera‐ ture simulation of forested riparian areas: I. Watershed model development. ASCE -

[17] Chinkuyu, A. J., T. Meixner, T. Gish, and C. Daughtry. 2004. The importance of seep‐ age zones in predicting soil moisture content and surface runoff using GLEAMS and

[18] Chinkuyu, A. J., and R. S. Kanwar. 2001. Predicting soil nitrate−nitrogen losses from incorporated poultry manure using GLEAMS model. Trans. ASAE, 44 (6): 1643−1650.

[19] Cronshey R. G., and Theurer F. G. 1998. AnnAGNPS-non point pollutant loading model. In Proceedings First Federal Interagency Hydrologic Modelling Conference.

[20] Dessu, S. B. and A. M. Melesse. 2012. Modelling the rainfall-runoff process of the Mara River basin using the Soil and Water Assessment Tool. Hydrological Processes,

nois Urban Hydrologic Model. Water Resources Research, 47: W08538.

Earth System Sciences, 15: 807–818.

EPA-600/R-93/174, U.S. EPA, Athens, GA, 30605

mental Protection Agency, Athens, GA.

Management, 26(10): 3023-3051.

ter conservation, 65 (6): 424-437.

Journal of Environ Engineering, 124:304–315

RZWQM. Trans of the ASAE, 47(2): 427−438.

Sheet, 3104, 4 p.

Las Vegas, NV.

DOI: 10.1002/hyp.9205.

## **Author details**

Prem B. Parajuli1\* and Ying Ouyang2

\*Address all correspondence to: pparajuli@abe.msstate.edu

1 Department of Agricultural and Biological Engineering at Mississippi State University, Mississippi State, USA

2 USDA-Forest Service Center for Bottomland Hardwoods Research, Mississippi State, USA

## **References**


[7] Betrie G. D., Y. A. Mohamed A. van Griensven and R. Srinivasan. 2011. Sediment management modelling in the Blue Nile Basin using SWAT model. Hydrology and Earth System Sciences, 15: 807–818.

**Acknowledgement**

**Author details**

Mississippi State, USA

**References**

Prem B. Parajuli1\* and Ying Ouyang2

The part of the case study research presented in this book chapter is based on work support‐ ed by the Special Research Initiatives (SRI) and Mississippi Agricultural and Forestry Ex‐

1 Department of Agricultural and Biological Engineering at Mississippi State University,

2 USDA-Forest Service Center for Bottomland Hardwoods Research, Mississippi State, USA

[1] Akter A., and Babel M. S. 2012. Hydrological modeling of the Mun River basin in

[2] Alam M. J., and Goodall J. L. 2012. Toward disentangling the effect of hydrologic and nitrogen source changes from 1992 to 2001 on incremental nitrogen yield in the con‐ tiguous United States. Water Resources Research, 48 (4), W04506, doi:

[3] Alarcon V. J., McAnally W. H.,and Pathak S. 2012. Comparison of two hydrodynamic models of Weeks Bay, Alabama. Computational science and its applications – ICC‐ SA, Lecture notes in Computer Science, 7334: 589-598, doi:

[4] Arnold J. G., Srinivasan R., Muttiah R. S., and Williams J. R. 1998. Large area hydro‐ logic modeling and assessment, Part I: model development. Journal of American Wa‐

[5] Baginska B., Milne-Home W., and Cornish P. S. 2003. Modelling nutrient transport in Currency Creek, NSW with AnnAGNPS and PEST. Environmental Modelling & Soft‐

[6] Bakhsh A., R. S. Kanwar D. B. Jaynes T. S. Colvin and L. R. Ahuja. 2000. Prediction of NO3−N losses with subsurface drainage water from manured and UAN−fertilized

periment Station (MAFES) at Mississippi State University.

74 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

\*Address all correspondence to: pparajuli@abe.msstate.edu

Thailand. Journal of Hydrology, 452-453: 232-246.

10.1029/2011WR010967 (in press).

10.1007/978-3-642-31075-1\_44.

ware, 18: 801–808.

ter Resources Association, 34(1): 73-89.

plots using GLEAMS. Trans. ASAE, 43 (1): 69−77.


[21] Donigian A. S. Jr., and Crawford N. H. 1976. Modeling Pesticides and Nutrients on Agricultural Lands. Environmental Research Laboratory, Athens, GA. EPA 600/2-7-76-043, 317 p

[33] Knisel W. G. 1980. CREAMS, a field scale model for chemicals, runoff and erosion from agricultural management systems. USDA Conservation Research Rept. No. 26,

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

77

[34] Leonard R. A., W. G. Knisel, and D. A. Still. 1987. GLEAMS: Groundwater loading effects of agricultural management systems. Trans. of the ASAE, 30:1403-1418.

[35] Licciardello F., D. A. Zema, S. M. Zimbone, and R. L. Bingner. 2007. Runoff and soil erosion evaluation by the AnnAGNPS model in a small Mediterranean watershed.

[36] Licciardello F., D. A. Zema, S. M. Zimbone, and R. L. Bingner. 2011. Runoff and soil erosion evaluation by the AnnAGNPS model in a small Mediterranean watershed.

[37] Mississippi Department of Environmental Quality (MDEQ), 2008. Sediment TMDL for the Yalobusha River Yazoo River Basin. PO Box 10385, Jackson, MS 39289-0385.

[38] Mississippi Department of Environmental Quality (MDEQ). 2010. Total Daily Maxi‐ mum Daily Load Program. Available at: http://www.deq.state.ms.us/MDEQ.nsf/ page/TWB\_Total\_Maximum\_Daily\_Load\_Section?OpenDocument. Accessed on

[39] Moriasi D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and Veith, T. L. 2007. Model evaluation guidelines for systematic quantification of accura‐

[40] National Climatic Data Center (NCDC). 2012. Locate weather observation station re‐ cord. Available at: http://www.ncdc.noaa.gov/oa/climate/stationlocator.html. Ac‐

[41] Neitsch S. L., Arnold, J. G., Kiniry, J. R., Williams, J. R., and King K. W. 2002. Soil and water assessment tool (SWAT), theoretical documentation, Blackland research center, grassland, soil and water research laboratory, agricultural research service: Temple,

[42] Neitsch S. L., Arnold J. G., Kiniry J. R., and Williams J. R. 2005. Soil and water assess‐ ment tool (SWAT), theoretical documentation, Blackland research center, grassland,

[43] Nejadhashemi A. P., Woznicki S. A. and Douglas‐Mankin K. R., 2011. Comparison of four models (STEPL, PLOAD, L‐THIA, AND SWAT) in simulating sediment, nitro‐ gen, and phosphorus loads and pollutant source areas. Trans. of the ASABE, 54(3):

[44] Ouyang Y., J. Higman and J. Hatten. 2012. Estimation of dynamic load of mercury in a river with BASINS-HSPF model. Journal of Soils and Sediments, 12(2): 207-216.

soil and water research laboratory, agricultural research service: Temple, TX.

cy in watershed simulations. Transactions of the ASABE, 50(3): 885-900.

U.S. Department of Agriculture: Washington D. C.

Trans. of the ASAE, 50, 1585-1593.

Trans. of the ASABE, 50(5): 1585-1593.

March 15, 2012.

cessed on April 14, 2012.

TX.

875-890.


[33] Knisel W. G. 1980. CREAMS, a field scale model for chemicals, runoff and erosion from agricultural management systems. USDA Conservation Research Rept. No. 26, U.S. Department of Agriculture: Washington D. C.

[21] Donigian A. S. Jr., and Crawford N. H. 1976. Modeling Pesticides and Nutrients on Agricultural Lands. Environmental Research Laboratory, Athens, GA. EPA

[22] Donogian A. S. Jr., Imhoff J. C., Bicknell B. R., Kittle J. I. 1984. Application guide for hydrological simulation program-FORTRAN (HSPF), EPA, Athens, GA.

[23] Dun S., J. Q. Wu, D. K. McCool, J. R. Frankenberger, and D. C. Flanagan. 2010. Im‐ proving frost-simulation sub-routines of the Water Erosion Prediction Project

[24] Flanagan D. C., J. E. Gilley, and T. G. Franti. 2007. Water Erosion Prediction Project (WEPP): development history, model capabilities, and future enhancements. Trans.

[25] Foltz R. B., W. J. Elliot, and N. S. Wagenbrenner. 2011. Soil erosion model predictions using parent material/soil texture-based parameters compared to using site-specific

[26] Gassman P. W., Reyes M. R., Green C. H., and Arnold, J. G. 2007. The Soil and Water Assessment Tool: historical development, applications, and future research direc‐

[27] Githui F., B. Selle and T. Thayalakumaran. 2012. Recharge estimation using remotely sensed evapotranspiration in an irrigated catchment in southeast Australia. Hydro‐

[28] Gong W., Wang Y., and Jia J. 2012. The effect of interacting downstream branches on saltwater intrusion in the Modaomen Estuary, China. Journal of Asian Earth Scien‐

[29] Gosain A. K., Rao S., Srinivasan R., and Reddy N. G. 2005. Return-flow assessment for irrigation command in the Palleru River basin using SWAT model. Hydrological

[30] Guedon, N. B., and Thomas, J. V. 2004. State of Mississippi Water Quality Assess‐ ment, section 305(b) Report 62. Mississippi Department of Environmental Quality,

[31] Hamrick, J. M. 1996. A User's Manual for the Environmental Fluid Dynamics Com‐ puter Code (EFDC), The College of William and Mary, Virginia Institute for Marine

[32] Hua L., Xiubin H., Yongping Y., and Hongwei N. 2012. Assessment of Runoff and Sediment Yields Using the AnnAGNPS Model in a Three-Gorge Watershed of China. International Journal of Environmental Research and Public Health, 9:1887-1907.

(WEPP) model. Trans. of the ASABE, 53(5): 1399-1411.

76 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

parameters. Trans. of the ASABE, 54(4): 1347-1356.

tions. Transactions of the ASABE, 50(4): 1211–1250.

600/2-7-76-043, 317 p

EPA-600/3-84-065.

of the ASABE, 50(5): 1603-1612.

logical Processes, 26: 1379-1389.

ces, 45: 223-238.

Jackson, MS.

Processes, 19: 673–682.

Sciences, Special Report 331.


[45] Parajuli P. B., Mankin K. R. and Barnes P. L. 2008. Applicability of Targeting Vegeta‐ tive Filter Strips to Abate Fecal Bacteria and Sediment Yield using SWAT. Agricultur‐ al Water Management, 95 (10): 1189-1200.

[58] Shields F. D. Jr., Cooper C. M., Testa III, S., and Ursic, M. E. 2008. Nutrient Transport in the Yazoo River Basin, Research Report 60. U.S. Dept of Agriculture Agricultural Research Service National Sedimentation Laboratory: Oxford. http://

Watershed-Scale Hydrological Modeling Methods and Applications

http://dx.doi.org/10.5772/53596

79

[59] Shirmohammadi A., B. Ulen, L. F. Bergstrom, and W. G. Knisel. 1998. Simulation of nitrogen and phosphorus leaching in a structured soil using GLEAMS and a new

[60] Sinha S., Liu X., and Garcia, M. H. 2012. Three-dimensional hydrodynamic modeling

[61] Smith R. A., Schwarz G. E., and Alexander R. B. 1997. Regional interpretation of wa‐ ter-quality monitoring data. Water Resources Research, 33 (12): 2781-2798.

[62] Talei A., and Chua L. H. C. 2012. Influence of lag time on event-based rainfall-runoff modeling using the data driven approach. Journal of Hydrology, 438-439: 223-233.

[63] U.S. Department of Agriculture (USDA). 2005. Soil data mart. Natural Resources Conservation Service. Available at: http://soildatamart.nrcs.usda.gov/Default.aspx.

[64] U.S. Department of Agriculture, National Agricultural Statistics Service (USDA/ NASS). 2010. The cropland data layer. Available at: http://www.nass.usda.gov/

[65] U.S. Environmental Protection Agency (US/EPA). 2010. National summary of im‐ paired waters and TMDL information. U.S. Environmental Protection Agency: Wash‐ ington, D.C. Available at: http://iaspub.epa.gov/waters10/attains\_nation\_cy.control?

[66] U.S. Geological Society (USGS). 1999. National elevation dataset. Available at: http://

[67] Vache K. B., Eilers J. M., and Santelmann M. V. 2002. Water quality modeling of alter‐ native agricultural scenarios in the U.S. Corn Belt. Journal of American Water Re‐

[68] Varanou E., Gkouvatsou E., Baltas E., and Mimikou M., 2002. Quantity and quality integrated catchment modeling under climate change with use of soil and water as‐ sessment tool model. ASCE Journal of Hydrological Engineering, 7(3): 228-244.

[69] Vellidis G., P. Barnes, D. D. Bosch, and A. M. Cathey. 2012. Mathematical simulation tools for developing dissolved oxygen TMDLs. Trans of the ASABE, 49(4):

[70] Williams, J. R., C. A. Jones and P. T. Dyke. 1984. A modeling approach to determin‐ ing the relationship between erosion and soil productivity. Trans. ASAE 27(1):

research/Cropland/SARS1a.htm. Accessed on April 29, 2012.

p\_report\_type=T. Accessed on June 28, 2012.

sources Association, 38(3): 773-787.

1003−1022.

129-144.

seamless.usgs.gov/index.php. Accessed April 29, 2012.

of the Chicago River, IL. Environmental Fluid Mechanics , Pages 1-24.

www.ars.usda.gov /SP2UserFiles/person/5120/NSLReport60.pdf.

sub-model. Trans. of the ASAE, 41(2): 353-360.

Accessed on May 23, 2012.


[58] Shields F. D. Jr., Cooper C. M., Testa III, S., and Ursic, M. E. 2008. Nutrient Transport in the Yazoo River Basin, Research Report 60. U.S. Dept of Agriculture Agricultural Research Service National Sedimentation Laboratory: Oxford. http:// www.ars.usda.gov /SP2UserFiles/person/5120/NSLReport60.pdf.

[45] Parajuli P. B., Mankin K. R. and Barnes P. L. 2008. Applicability of Targeting Vegeta‐ tive Filter Strips to Abate Fecal Bacteria and Sediment Yield using SWAT. Agricultur‐

[46] Parajuli P. B., Mankin K. R., and Barnes P. L. 2009. Source specific fecal bacteria mod‐ eling using soil and water assessment tool model. Bioresource Technology, 100 (2):

[47] Parajuli P. B. 2010a. Assessing sensitivity of hydrologic responses to climate change from forested watershed in Mississippi. Hydrological Processes, 24 (26): 3785-3797.

[48] Parajuli P. B. 2010b. Methods for Modeling Livestock and Human Sources of Nu‐

[49] Parajuli P. B. 2011. Effects of Spatial Heterogeneity on Hydrologic Responses at Wa‐

[50] Parajuli P. B. 2012. Evaluation of Spatial Variability on Hydrology and Nutrient Source Loads at Watershed Scale using a Modeling Approach. Hydrology Research,

[51] Polyakov V., Fares A., Kubo D., Jacobi J., and Smith C. 2007. Evaluation of a nonpoint source pollution model, AnnAGNPS, in a tropical watershed. Environmental

[52] Renard K. G., Foster G. R., Weesies G. A., McCool D. K., and Yoder D. C. 1997. Pre‐ dicting Soil Erosion by Water: A Guide to Conservation Planning with the Revised Universal Soil Loss Equation (RUSLE). U.S. Department of Agriculture, Agriculture

[53] Rolle K., Gitau M. W., Chen G., and Chauhan A. 2012. Assessing fecal coliform fate and transport in a coastal watershed using HSPF. Water Science and Technology,

[54] Saad D. A., Schwarz G. E., Robertson D. M., and Booth N. L. 2011. A Multi-Agency Nutrient Dataset Used to Estimate Loads, Improve Monitoring Design, and Calibrate Regional Nutrient SPARROW Models. Journal of the American Water Resources As‐

[55] Schwarz, G. E. 2008. A Preliminary SPARROW model of suspended sediment for the conterminous United States, U.S. Geological Survey Open-File Report 2008–1205, 8 p.

[56] Sharpley A. N. and J. R. Williams. 1990. EPIC—Erosion/Productivity Impact Calcula‐ tor: 1. Model Documentation. U.S. Department of Agriculture Technical Bulletin No.

[57] Sheshukov A. Y., Siebenmorgen C. B., and Douglas‐Mankin K. R. 2011. Seasonal and annual impacts of climate change on watershed response using an ensemble of glob‐

al climate models. Trans. of the ASABE, 54(6): 2209-2218.

trients at Watershed Scale. MAFES Research Report, 24(8): 1-8.

tershed Scale. Journal of Environmental Hydrology, 19(18): 1-18.

al Water Management, 95 (10): 1189-1200.

78 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

953-963.

doi:10.2166/nh.2012.013.

Handbook No. 703.

sociation, 47: 933-949.

66:1096-1102.

1768.

Modelling & Software, 22(11): 1617–1627.


[71] Williams J. R. and A. D. Nicks. 1985. SWRRB, a simulator for water resources in rural basins: an overview. In: D.G. DeCoursey (editor), Proc. of the Natural Resources Modeling Symp., Pingree Park, CO. USDA-ARS, ARS-30, pp. 17-22.

**Section 2**

**Contaminant Hydrology: Groundwater**


**Contaminant Hydrology: Groundwater**

[71] Williams J. R. and A. D. Nicks. 1985. SWRRB, a simulator for water resources in rural basins: an overview. In: D.G. DeCoursey (editor), Proc. of the Natural Resources

[72] Yuan Y., Bingner R. L., and Rebich R. A. 2001. Evaluation of AnnAGNPS on Missis‐

[73] Yuan Y, Dabney S, and Bingner R. L. 2002. Cost/benefit analysis of agricultural BMPs for sediment reduction in the Mississippi Delta. Journal of Soil and Water Conserva‐

[74] Yuan Y., Locke M. A., and Bingner R. L. 2008. Annualized Agricultural Non-Point Source model application for Mississippi Delta Beasley Lake watershed conservation

[75] Yuan Y., R. L. Bingner M. A. Locke F. D. Theurer, and J. Stafford. 2011. Assessment of Subsurface Drainage Management Practices to Reduce Nitrogen Loadings Using An‐

[76] Young R. A., Onstead C. A., Bosch D. D., and Anderson W. P. 1989. AGNPS: a non‐ point source pollution model for evaluating agricultural watersheds. Journal of Soil

[77] Zhang G. H., B. Y. Liu and X. C. Zhang. 2008. Applicability of WEPP sediment trans‐

[78] Zuercher B. W., D. C. Flanagan, and G. C. Heathman. 2011. Evaluation of the An‐ nAGNPS model for atrazine prediction in Northeast Indiana. Trans of the ASABE,

practices assessment. Journal of Soil and Water Conservation, 63(6): 542-551.

nAGNPS. Applied Engineering in Agriculture, 27(3): 335-344.

port equation to steep slopes. Trans of the ASABE, 51(5): 1675-1681.

Modeling Symp., Pingree Park, CO. USDA-ARS, ARS-30, pp. 17-22.

80 Current Perspectives in Contaminant Hydrology and Water Resources Sustainability

sippi Delta MSEA watersheds. Trans. of the ASAE, 44(5): 1183–1190.

tion, 57(5): 259–267.

54(3): 811-825.

Water Conservation, 44(2):168–173.

**Chapter 4**

**Arsenic in Groundwater: A Summary of Sources and the**

**Biogeochemical and Hydrogeologic Factors Affecting**

**1.1. World-wide occurrences of arsenic–contaminated groundwater – Forms and toxicity**


Arsenic (As) is a metalloid element (atomic number 33) with one naturally occurring isotope of atomic mass 75, and four oxidation states (-3, 0, +3, and +5) (Smedley and Kinniburgh, 2002). In the aqueous environment, the +3 and +5 oxidation states are most prevalent, as the

~4-10) (Smedley and Kinniburgh, 2002). In soils, arsine gases (containing As3-) may be gener‐

The different forms of As have different toxicities, with arsine gas being the most toxic form. Of the inorganic oxyanions, arsenite is considered more toxic than arsenate, and the organic (methylated) arsenic forms are considered least toxic (for a detailed discussion of toxicity issues, the reader is referred to Mandal and Suzuki (2002)). Arsenic is a global health concern due to its toxicity and the fact that it occurs at unhealthful levels in water supplies, particularly

Despite its use in medicines for nearly 2,500 years (Mandal and Suzuki, 2002; Cullen, 2008) As has long been recognized as a toxic and often lethal substance. Chronic exposure to As can cause harm to the human cardiovascular, dermal, gastrointestinal, hepatic, neurologi‐ cal, pulmonary, renal and respiratory systems (ATSDR, 2000) and reproductive system (Mandal and Suzuki, 2002). Research on health effects is summarized and discussed by

groundwater, in more than 70 countries (Ravenscroft et al., 2009) on six continents.

at pH ~9-11) and arsenate (H2AsO4

© 2013 Barringer and Reilly; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2013 Barringer and Reilly; licensee InTech. This is a paper distributed under the terms of the Creative Commons


2- at pH

**Arsenic Occurrence and Mobility**

Additional information is available at the end of the chapter

Julia L. Barringer and Pamela A. Reilly

http://dx.doi.org/10.5772/55354

oxyanions arsenite (H3AsO3 or H2AsO3

**1.2. Health effects and standards**

ated by fungi and other organisms (Woolson, 1977).

**1. Introduction**
