J. G. Arnold

Soil and Water Assessment Tool (SWAT) Model: Current Developments and Applications

This article introduces a special collection of 20 research articles that present current developments and applications of the Soil and Water Assessment Tool (SWAT). The first objective is to review and introduce the research addressed within this special collection. The second objective is to summarize and synthesize the model performance statistics and parameters published in these articles to provide a succinct guide to complement a previous SWAT model summary. Recent SWAT developments in landscape representation, stream routing, and soil P dynamics are presented in this collection. Numerous critical applications of the SWAT model were conducted across a variety of landscape scales, climatic and physiographic regions, and pollutant sources. In this article, model performance in terms of coefficient of determination, Nash-Sutcliffe efficiency, and percent bias across all the studies is summarized and found to be satisfactory or better in all cases. These results are then compiled with a previous synthesis of results to generate a comprehensive assessment of SWAT. Model parameters used to calibrate the model for streamflow, sediment, N, and P in numerous studies are also summarized. This collection demonstrates that research in development and application of the SWAT model and associated tools continues to grow internationally in a wide range of settings and applications.

SWAT Ungauged: Hydrological Budget and Crop Yield Predictions in the Upper Mississippi River Basin

Physically based, distributed hydrologic models are increasingly used in assessments of water resources, best management practices, and climate and land use changes. Model performance evaluation in ungauged basins is an important research topic. In this study, we propose a framework for developing Soil and Water Assessment Tool (SWAT) input data, including hydrography, terrain, land use, soil, tile, weather, and management practices, for the Upper Mississippi River basin (UMRB). We also present a performance evaluation of SWAT hydrologic budget and crop yield simulations in the UMRB without calibration. The uncalibrated SWAT model ably predicts annual streamflow at 11 USGS gauges and crop yield at a four-digit hydrologic unit code (HUC) scale. For monthly streamflow simulation, the performance of SWAT is marginally poor compared with that of annual flow, which may be due to incomplete information about reservoirs and dams within the UMRB. Further validation shows that SWAT can predict base flow contribution ratio reasonably well. Compared with three calibrated SWAT models developed in previous studies of the entire UMRB, the uncalibrated SWAT model presented here can provide similar results. Overall, the SWAT model can provide satisfactory predictions on hydrologic budget and crop yield in the UMRB without calibration. The results emphasize the importance and prospects of using accurate spatial input data for the physically based SWAT model. This study also examines biofuel-biomass production by simulating all agricultural lands with switchgrass, producing satisfactory results in estimating biomass availability for biofuel production.

APPLICATION OF A WATERSHED MODEL TO EVALUATE MANAGEMENT EFFECTS ON POINT AND NONPOINT SOURCE POLLUTION

A Total Maximum Daily Load (TMDL) program has been initiated in the North Bosque River Watershed in Texas, USA, where point and nonpoint sources of pollution are of a concern. The Soil and Water Assessment Tool (SWAT), which had been validated for flow and sediment and nutrient transport, was applied to quantify the effects of Best Management Practices (BMPs) related to dairy manure management and municipal wastewater treatment plant effluent. Results are presented for the period from 1960 through 1998 for three sites along the North Bosque River. Results are presented as annual time–weighted concentrations (average of the daily load divided by daily flow over a year) and annual flow–weighted concentrations (total cumulative load divided by total cumulative flow over a year). The wastewater treatment plant BMPs resulted in greater improvement in time–weighted instream soluble phosphorus concentrations than dairy BMPs. On the other hand, dairy BMPs made greater differences in flow–weighted concentrations. This study showed that SWAT could be a useful tool for studying the effects of alternative management scenarios for pollution control from point and nonpoint sources in large watersheds.

HYDROLOGIC EVALUATION OF THE SOIL AND WATER ASSESSMENT TOOL FOR A LARGE TILE-DRAINED WATERSHED IN IOWA

The presence of subsurface tile drainage systems can facilitate nutrient and pesticide transport, thereby contributing to environmental pollution. The Soil and Water Assessment Tool (SWAT) water quality model is designed to assess nonpoint and point source pollution and was recently modified for tile drainage. Over 25% of the nations cropland required improved drainage. In this study, the models ability to validate the tile drainage component is evaluated with nine years of hydrologic monitoring data collected from the South Fork watershed in Iowa, since about 80% of this watershed is tile drained. This watershed is a Conservation Effects Assessment Program benchmark watershed and typifies one of the more intensively managed agricultural areas in the Midwest. Comparison of measured and predicted values demonstrated that inclusion of the tile drainage system is imperative for obtaining a realistic watershed water balance. Two calibration/validation scenarios tested if the results differed in how the data set was divided. The optimum scenario results for the simulated monthly and daily flows had Nash-Sutcliffe efficiency (ENS) values during the calibration/validation (1995-1998/1999-2004) periods of 0.9/0.7 and 0.5/0.4, respectively. The second scenario results for the simulated monthly and daily flows had ENS values during the calibration/validation (1995-2000/2001-2004) periods of 0.8/0.5 and 0.7/0.2, respectively. The optimum scenario reflects the distribution of peak rainfall events represented in both the calibration and validation periods. The year 2000, being extremely dry, negatively impacted both the calibration and validation results. Each water budget component of the model gave reasonable output, which reveals that this model can be used for the assessment of tile drainage with its associated practices. Water yield results were significantly different for the simulations with and without the tile flow component (25.1% and 16.9%, expressed as a percent of precipitation). The results suggest that the SWAT2005 version modified for tile drainage is a promising tool to evaluate streamflow in tile-drained regions when the calibration period contains streamflows representing a wide range of rainfall events.

Soil and Water Assessment Tool (SWAT) Hydrologic/Water Quality Model: Extended Capability and Wider Adoption

This article introduces a special collection of 16 research articles on new developments and applications of the Soil and Water Assessment Tool (SWAT) to address various environmental issues at a range of geographic and temporal scales. Highlights include addition of a subdaily erosion and sediment transport algorithm, a biozone module, and a new algorithm for shallow water table depth. Model applications include climate change impact assessments, model adaptation to regional environmental conditions, watershed-scale soil erosion assessments, and linkages to other models. A summary of reported model performance indicates that 85% of daily flow calibration statistics reported in this collection were satisfactory or better, with very good performance in four of the 20 calibration results and in three of the 19 validation results. Details of reported model parameters for calibration of flow and water quality constituents are provided for other SWAT modelers. This collection builds upon a previous ASABE 2010 SWAT Special Collection, demonstrating continued developments to enhance SWATs capabilities and highlighting SWATs continued expansion in international applications, especially in Asia.

ESTIMATION OF LONG-TERM SOIL MOISTURE USING A DISTRIBUTED PARAMETER HYDROLOGIC MODEL AND VERIFICATION USING REMOTELY SENSED DATA

Soil moisture is an important hydrologic variable that controls various land surface processes. In spite of its importance to agriculture and drought monitoring, soil moisture information is not widely available on a regional scale. However, long-term soil moisture information is essential for agricultural drought monitoring and crop yield prediction. The hydrologic model Soil and Water Assessment Tool (SWAT) was used to develop a long-term record of soil water at a fine spatial (16 km 2 ) and temporal (weekly) resolution from historical weather data. The model was calibrated and validated using stream flow data. However, stream flow accounts for only a small fraction of the hydrologic water balance. Due to the lack of measured evapotranspiration or soil moisture data, the simulated soil water was evaluated in terms of vegetation response, using 16 years of normalized difference vegetation index (NDVI) derived from NOAA-AVHRR satellite data. The simulated soil water was well-correlated with NDVI (r as high as 0.8 during certain years) for agriculture and pasture land use types, during the active growing season April-September, indicating that the model performed well in simulating the soil water. The study provides a framework for using remotely sensed NDVI to verify the soil moisture simulated by hydrologic models in the absence of auxiliary measured data on ET and soil moisture, as opposed to just the traditional stream flow calibration and validation.

Evaluating hydrology of the Soil and Water Assessment Tool (SWAT) with new tile drain equations

Although subsurface drainage is a water management system widely used to maximize crop production in regions with seasonal high water tables, such as the midwestern United States, it is also a major source of nutrients into water bodies. Recently, physically based Hooghoudt and Kirkham tile drain equations were incorporated into the Soil and Water Assessment Tool (SWAT) model (herein referred to as Modified SWAT) as alternative tile flow simulation methods and a tool to design cost-effective and environment-friendly tile drain water management systems. The goal of this study was to determine a range of values for the new tile drain parameters and to use measured streamflow data from the South Fork Watershed (SFW) in Iowa to evaluate the capability of the Modified SWAT to simulate water balance components for this tile-drained watershed. This was accomplished by reviewing literature of tile drainage studies and by comparing measured streamflow with that predicted by the Modified SWAT using the Nash-Sutcliffe efficiency (NSE) and percent bias (PBIAS [%]) statistical methods in addition to hydrographs. During the calibration period, the Modified SWAT simulated streamflow very well (monthly NSE = 0.85 and PBIAS = ±2.3%). During the validation period, the Modified SWAT model simulated streamflow well (monthly NSE = 0.70 and PBIAS = ±2.5%). Simulated water balance results indicated that the soil water with tile drainage (260 mm [10 in]) was significantly (p-value = 0.00) lower than soil water without tile drainage (355 mm [14 in]), while streamflow with (205 mm [8 in]) tile drainage was significantly (p-value = 0.03) greater than streamflow without (128 mm[5 in]) tile drainage. This shows that the Hooghoudt steady-state and Kirkham tile drain equations are potential alternative tile flow simulation methods and tile drainage design tools in SWAT.

An integrated modeling approach for estimating the water quality benefits of conservation practices at the river basin scale.

The USDA initiated the Conservation Effects Assessment Project (CEAP) to quantify the environmental benefits of conservation practices at regional and national scales. For this assessment, a sampling and modeling approach is used. This paper provides a technical overview of the modeling approach used in CEAP cropland assessment to estimate the off-site water quality benefits of conservation practices using the Ohio River Basin (ORB) as an example. The modeling approach uses a farm-scale model, Agricultural Policy Environmental Extender (APEX), and a watershed scale model (the Soil and Water Assessment Tool [SWAT]) and databases in the Hydrologic Unit Modeling for the United States system. Databases of land use, soils, land use management, topography, weather, point sources, and atmospheric depositions were developed to derive model inputs. APEX simulates the cultivated cropland, Conserve Reserve Program land, and the practices implemented on them, whereas SWAT simulates the noncultivated land (e.g., pasture, range, urban, and forest) and point sources. Simulation results from APEX are input into SWAT. SWAT routes all sources, including APEXs, to the basin outlet through each eight-digit watershed. Each basin is calibrated for stream flow, sediment, and nutrient loads at multiple gaging sites and turned in for simulating the effects of conservation practice scenarios on water quality. Results indicate that sediment, nitrogen, and phosphorus loads delivered to the Mississippi River from ORB could be reduced by 16, 15, and 23%, respectively, due to current conservation practices. Modeling tools are useful to provide science-based information for assessing existing conservation programs, developing future programs, and developing insights on load reductions necessary for hypoxia in the Gulf of Mexico.

Detection of Overparameterization and Overfitting in an Automatic Calibration of SWAT

Distributed hydrologic models based on small-scale physical processes tend to have a large number of parameters to represent spatial heterogeneity. This characteristic requires the use of a large number of parameters in model calibration. It is a common view that calibration with a large number parameters produces overparameterization and overfitting. Recent work using prior information, spatial information, and constraints on parameters for regularization of the calibration problem has improved model predictions using a few dozen parameters. We demonstrate that the Soil and Water Assessment Tool (SWAT) and the information associated with a SWAT watershed setup provide a regularized problem with many of recently published regularization techniques already utilized in SWAT. Our hypothesis is that the Soil and Water Assessment Tool (SWAT) regularizes the inverse problem so that a stable solution can be obtained for calibration of SWAT using a very large number of parameters, where very large means up to 10,000 calibration parameters. In this study, a two-objective calibration genetic algorithm based on a non-dominated sorting genetic algorithm (NSGA-II) was used to calibrate the Blue River basin in Oklahoma. We introduce the use of intermediate solutions found by the genetic algorithm to test identification of calibration parameters and diagnose model overfitting. Defining identification as the capability of a model to constrain the estimation of parameters, we introduced a method for statistically testing for changes from the initial uniform distribution of each parameter. We found that all 4,198 parameters used to calculate the Blue River SWAT model were identified. Diagnostic comparisons of goodness-of-fit measures for the calibration and validation periods provided strong evidence that the model was not overfitted.

Soil and Water Assessment Tool Hydrologic and Water Quality Evaluation of Poultry Litter Application to Small-Scale Subwatersheds in Texas

The application of poultry litter to agricultural land has become a topic of interest for policy makers due to public concern about its effects on water quality. The Soil and Water Assessment Tool (SWAT) version 2005 is designed to assess nonpoint and point sources of pollution. In this study, six subwatersheds in Texas (HUC-8; 12070101) are used to evaluate the models ability to simulate water quality at a small scale. Each of these subwatersheds randomly received poultry litter rates of 0.0 to 13.4 Mg ha -1 . Monthly and daily data from 2002 were used for calibration purposes, while 2000, 2001, 2003, and 2004 were used for validation. The SCS runoff curve number for moisture condition II (CN2) and the soil evaporation compensation factor (ESCO) parameters were found to be more sensitive than the surface runoff lag time (SURLAG) and initial soil water content expressed as a fraction of field capacity (FFCB). The monthly and daily runoff model simulations for the six subwatersheds resulted in calibration Nash-Sutcliffe efficiency (NSE) values of 0.59 and 0.53 and validation NSE values 0.82 and 0.80, respectively. The monthly and daily R 2 runoff values for the six subwatersheds resulted in calibration values of at least 0.60 and 0.53 and validation R 2 values of 0.86 and 0.81, respectively. The observed trends included SWATs overestimation of runoff in the dry periods and underestimation in the wet periods. The monthly NSE and R 2 values for sediment and nutrient losses were generally above 0.4 and 0.5, respectively. Paired t-tests for the monthly manually adjusted parameter simulation of sediment, organic N and P, NO3-N, and soluble P for the 2000-2004 period losses showed that their respective SWAT means were not significantly different from the measured values ( = 0.05), except for NO3-N losses for the Y10 subwatershed (p-value 0.042). The control subwatersheds measured and simulated water quality results were significantly different ( = 0.05) from the treated subwatersheds, most likely due to the amount of inorganic N present. Almost all of the subwatersheds that had poultry litter applied resulted in higher sediment, organic N, organic P, and soluble P losses than the control subwatershed upon averaging the monthly validation values. High NO3-N losses may have been a function of poultry litter and commercial fertilizers being applied before a large rainfall event occurred. The subwatersheds that received smaller amounts of commercial fertilizer and/or poultry litter lost more sediment, organic N, and organic P than the subwatersheds that received the higher litter and/or fertilizer treatments. Overall, the SWAT simulated the hydrology and the water quality constituents at the subwatershed scale more adequately when all of the data were used to simulate the model, as evidenced by statistical measures.