To better understand the Agricultural Policy Environmental Extender (APEX) model's strengths and limitations in simulating streamflow and water quality, we evaluated its performance under a diverse range of climatic, topographic, soil, cover, and land management conditions using three parameter settings: best professional judgment (BPJ), partially calibrated, and fully calibrated. A total of 18 calibration parameters governing streamflow, crop yield, and sediment, N, and P were adjusted. Hydrologic and water quality responses were simulated on ten USDA-ARS watersheds with heterogeneous forested, pasture/range, and/or corn-soybean cropping systems. Based on percent bias and Nash-Sutcliffe coefficient of efficiency test statistics, 0%, 0%, 15%, and 9% of the data sets were considered satisfactory or better in simulating monthly streamflow, sediment, N, and P, respectively, in the BPJ mode. However, when used with the fully calibrated settings, 73%, 43%, 38%, and 27% of the data sets were satisfactory or better. Findings indicate that when fully calibrated, APEX estimates streamflow very well, sediment and nitrogen moderately well, and phosphorus marginally well at a monthly time scale. However, when using the BPJ approach, APEX does not estimate watershed-level streamflow, sediment, or nutrients accurately at a monthly time step. Additionally, APEX lacks an element of robustness in simulating streamflow and water quality constituents when applied to the same watershed with a different period of record or to a nearby watershed with a similar period of record. Based on these findings, it is recommended that users exercise caution when employing APEX with the BPJ approach or in validation mode at watershed scales.

This study employed the Soil and Water Assessment Tool (SWAT) to evaluate the impacts of projected future climate change scenarios on water balance, runoff, sediment, total nitrogen (N), and total phosphorus (P) at the field scale for four locations in the Heartland region: Sioux City (Iowa) and Columbus, Mullen, and Harrison (Nebraska). A conventional two-year corn-soybean rotation was assumed to be grown on each field. All fields were simulated identically in terms of topographic and cover/land management conditions. Model inputs for the fields differed in only three ways: the forcing conditions for existing and future climatic scenarios (SRES A2, A1B, and B1), soil and aquifer properties, and calibrated parameters at each location. Model simulations indicate that for the Columbus and Sioux City sites, where current average annual precipitation is about 740 and 650 mm, respectively, substantial increases in runoff and pollutant loadings from a corn-soybean crop rotation are projected to occur during the spring under future climate scenarios in comparison to existing conditions. At the Sioux City site, for example, increases in runoff of 213%, 124%, and 128% during the month of May are projected for the A2, A1B, and B1 scenarios, respectively, in comparison to the baseline condition. Very large increases in attendant sediment and nutrient losses are also projected for that month at the Sioux City site. Considerably greater attention in coming years will therefore likely be necessary to devise best management practices and adaptation strategies that can be effectively employed to conserve soil and water resources and to protect streams and receiving waters from the harmful effects of higher pollutant loadings. At the Harrison site, where average annual precipitation is less than 450 mm, increases in average annual evapotranspiration of 29, 31, and 46 mm under the A2, A1B, and B1 future climate scenarios are projected to occur for a corn-soybean crop. Relative to the baseline at this site, water requirements are projected to be 37, 39, and 32 mm greater, respectively, for the peak irrigation month of July under the A2, A1B, and B1 scenarios. For regions in western Nebraska with similar or lesser precipitation levels, these anticipated changes could exacerbate already existing challenges for agricultural producers who primarily rely upon groundwater for irrigation. SWAT was employed to simulate the impact of four best management practices (BMPs) on changes in sediment, total N, and total P under the baseline and future climate change scenarios for each site. These four treatments included conversion of the corn-soybean rotation to pasture, switchgrass, and no-till, and implementation of a 10 m wide edge-of-field buffer strip. At all four sites, the pasture and switchgrass BMPs reduced sediment and total P yields by 97% to 99% in comparison to the corn-soybean cover crop. Each of the BMP treatments employed in this study appears to hold promise in providing potential reductions in sediment and nutrients for the two eastern field sites. However, further analyses are needed to not only assess the impact of other types of BMPs, but their cost effectiveness and sustainability as well. Model simulations suggest that, for the Harrison site, moderate decreases in sediment, total N, and total P are projected to occur for the no-till BMP and modest decreases for the 10 m buffer strip BMP. Model simulations also suggest that, of the four BMP types, only the pasture and switchgrass treatments appear to provide appreciable reductions in sediment and nutrients at the Mullen site.

Federal and state agencies across the U.S. are currently tasked with Total Maximum Daily Load (TMDL) development to ensure compliance with the Clean Water Act of 1972. In the northwestern part of the country, the TMDL effort is particularly challenging due to the complicated nature of expansive forested watersheds, steep mountainous topography, and orographic precipitation. This is especially true for sediment, which is a primary pollutant of concern. Modeling, in combination with field source assessments, has commonly been used to estimate watershed sediment yields and associated source contributions. However even with widespread use of these methods, little has been done to evaluate the prediction performance of modeling tools in forested mountainous re ions. The purpose of this article is to present an eight-year simulation period (1985-1992) for the 1,709 km(2) Lamar River watershed in Yellowstone National Park where simulated loads from the Soil and Water Assessment Tool (SWAT) were compared with observed suspended sediment discharge data. Based on Nash-Sutcliffe model efficiencies of > 0.81 and > 0.86 for daily and monthly streamflow, and > 0.51 and > 0.78 for sediment, our findings suggest that SWAT is suitable for simulation of sediment in mountainous and snowmelt-dominated terrain. Two ancillary lines of evidence were used to support this conclusion: (1) a comparison of simulated landscape sediment yields with that of regional literature studies, and (2) confirmation of simulated landscape and bank erosion source contributions with that of ratios established using radionuclide tracers.

The USDA-ARS Conservation Effects Assessment Project (CEAP) calls for improved understanding of the strengths and weaknesses of watershed-scale water quality models under a range of climatic, soil, topographic, and land use conditions. Assessing simulation model parameter sensitivity helps establish feasible parameter ranges, distinguish among parameters having regional versus universal interactions, and ensure that one model process does not compensate for another due to poor parameter settings. The Soil and Water Assessment Tool (SWAT) parameter sensitivity and autocalibration module was tested on two northern and three southern USDA-ARS experimental watersheds. These previously calibrated watersheds represent a range of climatic, physiographic, and land use conditions present in the U.S. Sixteen parameters that govern basin, snow accumulation/melt, surface, and subsurface response in the model were evaluated. Parameters governing surface runoff due to rainfall! were found most sensitive overall, while parameters governing groundwater were the least sensitive. Surface runoff parameters were found most sensitive for areas with high evaporation rates and localized thunderstorms. Parameters from all categories were important in areas where precipitation includes both rainfall and snowfall. Differences in model performance were noticeable on a climatic basis; SWAT generally predicted streamflow with less uncertainty in humid climates than in arid or semi-arid climates. Study findings can be used to determine appropriate parameter ranges for ungauged watersheds of similar characteristics.

Watershed models are powerful tools for simulating the effect of watershed processes and management on soil and water resources. However, no comprehensive guidance is available to facilitate model evaluation in terms of the accuracy of simulated data compared to measured flow and constituent values. Thus, the objectives of this research were to: (1) determine recommended model evaluation techniques (statistical and graphical), (2) review reported ranges of values and corresponding performance ratings for the recommended statistics, and (3) establish guidelines for model evaluation based on the review results and project-specific considerations; all of these objectives focus on simulation of streamflow and transport of sediment and nutrients. These objectives were achieved with a thorough review of relevant literature on model application and recommended model evaluation methods. Based on this analysis, we recommend that three quantitative statistics, Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS), and ratio of the root mean square error to the standard deviation of measured data (RSR), in addition to the graphical techniques, be used in model evaluation. The following model evaluation performance ratings were established for each recommended statistic. In general, model simulation can be judged as satisfactory if NSE > 0.50 and RSR < 0.70, and if PBIAS + 25% for streamflow, PBIAS + 55% for sediment, and PBIAS + 70% for N and P. For PBIAS, constituent-specific performance ratings were determined based on uncertainty of measured data. Additional considerations related to model evaluation guidelines are also discussed. These considerations include: single-event simulation, quality and quantity of measured data, model calibration procedure, evaluation time step, and project scope and magnitude. A case study illustrating the application of the model evaluation guidelines is also provided.

Watershed computer models such as the Soil and Water Assessment Tool (SWAT) contain parameters that describe watershed properties such as vegetative cover, soil characteristics, or landscape features. For investigations that involve changes in land cover or land management on agricultural lands, proper adjustment of these parameters is important not only for runoff estimation, but also for the simulation of sediment, nutrients, and other pollutants. However, these parameters may only be known for a few small, homogeneous areas, and the usefulness of such parameters in calibrating the runoff response for a watershed scale model such as SWAT is not well documented. The objective of this study was to determine if model parameters that govern the surface runoff response in SWAT that were calibrated from rain-fed unit source area watersheds could be scaled up to provide accurate runoff simulations at a watershed scale. Model testing was conducted on four unit source area watersheds that consisted of homogeneous Bermuda grass, pasture, and winter wheat land cover types and three larger subwatersheds of the Little Washita River Experimental Watershed in southwestern Oklahoma. Data from the unit source area watersheds were used to calibrate parameters in SWAT that govern only the surface runoff output from the model. These parameter values were extended to the larger, 160 km(2) (61.9 mi(2)) subwatershed 526, and model simulations were then evaluated by examining both the surface runoff and total water yield response of the model. Simulation results from the unit source area watersheds suggest that the soil evaporation compensation factor (ESCO) in SWAT not only reflects soil field conditions for which it was intended to describe, but the impact of land management conditions on surface runoff response as well. Findings from this research indicate that if a value of ESCO that was calibrated from the unit source area watershed data for winter wheat was applied at the watershed scale, it would lead to model simulations that give a surface runoff to total runoff fraction that is more than 15% too high. Due to uncertainties in relating ESCO to soil and land management properties, results of this study suggest that runoff data from unit source area watersheds may be best suited for calibrating infiltration functions or verifying values of the runoff curve number for watershed simulations.

An investigation was conducted to evaluate strengths and limitations of manual calibration and the existing autocalibration tool in the watershed-scale model referred to as the Soil and Water Assessment Tool (SWAT). Performance of the model was tested on the Little River Experimental Watershed in Georgia and the Little Washita River Experimental Watershed in Oklahoma, both USDA-ARS watersheds. A long record of multi-gauge streamflow data on each of the watersheds was used for model calibration and validation. Model performance of the streamflow response in SWAT was assessed using a six-parameter manual calibration based on daily mass balance and visual inspection of hydrographs and duration of daily flow curves, a six-parameter autocalibration method based on the daily sum of squares of the residuals after ranking objective function (referred to as SSQRauto6), a six-parameter method based on the daily sum of squares of residuals (SSQauto6), and an eleven-parameter method based on the daily sum of square of residuals (SSQauto11). Results show that for both watersheds, manual calibration generally outperformed the autocalibration methods based on percent bias (PBIAS) and simulation of the range in magnitude of daily flows. For the calibration period on Little River subwatershed F, PBIAS was 0.0%, -24.0%, -21.5%, and +29.0% for the manual, SSQRauto6, SSQauto6, and SSQauto11 methods, respectively. Based on the coefficient of efficiency (NSE), the SSQauto6 and SSQauto11 methods gave substantially better results than manual calibration on the Little River watershed. On the Little Washita watershed, however, the manual approach generally outperformed the automated methods, based on the NSE error statistic. Results of this study suggest that the autocalibration option in SWAT provides a powerful, labor-saving tool that can be used to substantially reduce the frustration and uncertainty that often characterize manual calibrations. If used in combination with a manual approach, the autocalibration tool shows promising results in providing initial estimates for model parameters. To maintain mass balance and adequately represent the range in magnitude of output variables, manual adjustments may be necessary following autocalibration. Caution must also be exercised in utilizing the autocalibration tool so that the selection of initial lower and upper ranges in the parameters results in calibrated values that are representative of watershed conditions.

Flood retarding structures (FRSs) represent one of the most effective methods for reducing damages caused by flooding and sedimentation from agricultural land. The impacts of these structures on streamflow and sediment regime and their effectiveness in reducing watershed floods and associated soil losses under dry, average, and wet climatic conditions were investigated in this study. The Soil and Water Assessment Tool (SWAT) was used to determine differences in streamflow and sediment characteristics with and without flood retarding structures under varying climatic conditions on two subwatersheds of the Little Washita River Experimental Watershed (LWREW) in Southwestern Oklahoma. Differences in the reduction of sediment yield between the two subwatersheds were attributed to such factors as the combined sediment storage capacity of the impoundment structures, their distance from the watershed outlet, and the percent of watershed area controlled by the structures. Model simulations from this study confirm the importance of the flood abatement program in reducing flooding and soil losses from agricultural land, but also show that low flow conditions may be exacerbated by the FRSs.

ABSTRACT: Precipitation and streamflow data from three nested subwatersheds within the Little Washita River Experimental Watershed (LWREW) in southwestern Oklahoma were used to evaluate the capabilities of the Soil and Water Assessment Tool (SWAT) to predict streamflow under varying climatic conditions. Eight years of precipitation and streamflow data were used to calibrate parameters in the model, and 15 years of data were used for model validation. SWAT was calibrated on the smallest and largest sub‐watersheds for a wetter than average period of record. The model was then validated on a third subwatershed for a range in climatic conditions that included dry, average, and wet periods. Calibration of the model involved a multistep approach. A preliminary calibration was conducted to estimate model parameters so that measured versus simulated yearly and monthly runoff were in agreement for the respective calibration periods. Model parameters were then fine tuned based on a visual inspection of daily hydrographs and flow frequency curves. Calibration on a daily basis resulted in higher baseflows and lower peak runoff rates than were obtained in the preliminary calibration. Test results show that once the model was calibrated for wet climatic conditions, it did a good job in predicting streamflow responses over wet, average, and dry climatic conditions selected for model validation. Monthly coefficients of efficiencies were 0.65, 0.86, and 0.45 for the dry, average, and wet validation periods, respectively. Results of this investigation indicate that once calibrated, SWAT is capable of providing adequate simulations for hydrologic investigations related to the impact of climate variations on water resources of the LWREW.

ABSTRACT: Following the devastating floods of the 1940s, thousands of flood retarding structures were constructed in the Great Plains. The impacts of these structures on streamflow characteristics and their effectiveness in reducing floods under dry, average, and wet climatic conditions were investigated in this study. The setting for the study was a 160 km2 (61.9 mi2) experimental watershed in southwestern Oklahoma that contained 13 flood retarding structures, which controlled 65% of the drainage area. Thirty-three years of precipitation and eight years of streamflow data, in conjunction with computer simulations, were used to evaluate changes due to the flood retarding structures in annual, monthly, and daily streamflow characteristics at the outlet of the watershed. Simulation results indicate that installation of the flood retarding structures leads to a decrease in average annual streamflow of about 3%, which was attributed to an increase in average annual evaporation of this same amount due to the free water surface of the reservoirs created by the flood retarding structures. On a monthly time scale, the simulation results showed that under dry climatic conditions, the flood retarding structures caused a reduction in mean streamflow for all months of the year. On the other hand, during average and wet climatic conditions, changes in mean streamflow varied by month with May and October showing the greatest decreases, respectively. As expected, the greatest impact of the flood retaiding structures was on daily flow characteristics and maximum daily flows. Noticeable differences include the reduction of maximum daily discharges on the day of a storm event and the increase in streamflow from principal spillway releases in the days immediately following storm events. Annual maximum daily discharges were reduced by about 33%, with the reduction of the 5 and 10 year maximum daily flows being 23.8 to 14.6 cms (840 to 515 cfs) (39%) and from 36.0 to 20.6 cms (1,270 to 727 cfs) (43%), respectively. This study demonstrates the effectiveness of the flood retarding structures in reducing annual peak runoff events, thereby reducing flooding and related property damage. However, the flood retarding structures also reduce low streamflow values by trapping baseflow runoff from drainage areas above the impoundment structures. From a water quality and stream habitat preservation point of view, maintenance of a minimum baseflow can be critical.

Strengths and limitations of hydrologic simulation models are used as criteria for selecting a particular model for a given water resources application. The performance of the Soil and Water Assessment Tool (SWAT) and the Hydrologic Simulation Program-Fortran (HSPF) continuous simulation models was compared on eight nested agricultural watersheds within the Little Washita River Experimental Watershed (LWREW) and two agricultural watersheds adjacent to the LWREW within the Washita River Basin in southwestern Oklahoma. Two subwatersheds within the LWREW were first used to calibrate parameters in both models for a "wetter than average" period of record. Both models were then applied to six other subwatersheds within the LWREW and the two adjacent watersheds, under varying climatic conditions. Three quantitative and two qualitative evaluation criteria were used to assess streamflow simulated by SWAT and HSPF: computation of (1) deviation of streamflow volume, (2) coefficient of efficiency, and (3) prediction efficiency and visual inspection of (4) hydrographs and (5) flow duration curves. A comparison of model performance showed that while HSPF performed better on the watersheds used for calibration, SWAT gave better results on the validation watersheds. On one of the validation subwatersheds adjacent to the LWREW, values of deviation of streamflow volume were -38.7%, -13.3%, and -1.3% for SWAT and -64.3%, -81.1%, and -8.2% for HSPF under "much dryer than average," "dryer than average," and "near average" climatic conditions, respectively. Differences in model performance were mainly attributed to the runoff production mechanism in the two models. Results of this study showed that SWAT exhibited an element of robustness in that it gave more consistent results than HSPF in estimating streamflow for agricultural watersheds under various climatic conditions. SWAT may therefore be better suited for investigating the long-term impacts of climate variability on surface-water resources.

The National Weather Service (NWS) operates a network of Doppler-radar stations (NEXRAD, WSR-88D) that produce hourly-rainfall estimates, at approximately 4-km 2 resolution, with nominal coverage of 96% of the conterminous US. Utilization of these data by the NWS is primarily for the detection and modeling of extreme-weather events. Radar-precipitation estimates were compared with gauge estimates at six ARS watershed-research locations in Idaho, Arizona, Oklahoma, Georgia and Mississippi to evaluate the utility of these data for hydrologic and natural resources modeling applications. Radar precipitation estimates underestimated gauge readings for all locations except Tucson. In all cases, the total number of hours with measured-radar precipitation was much less than hours containing gauge-precipitation estimates. Additional modification of NWS precipitation processing procedures will be necessary to improve accessibility and utility of these data for hydrologic and natural-resource modeling applications.