Search

HighlightsDrainage water recycling captures and stores agricultural drainage water for reuse as supplemental irrigation.Drainage water recycling can both increase crop production and benefit downstream water quality.Depending on management, drainage water recycling can also provide other complementary benefits.Research needs to advance drainage water recycling are presented and discussed. Keywords: Drainage water quality, Drainage water reuse, Subsurface drainage, Supplemental irrigation, Agricultural resiliency.

HighlightsA modeling framework for drainage water recycling (DWR) was developed to estimate irrigation and water quality benefits.Global sensitivity analysis was used to identify most and least influential input parameters affecting model outputs.Parameters controlling total available water had the most influence on applied irrigation and captured tile drain flow.The modeling framework and sensitivity results were used to develop an open-source, online tool for evaluating DWR.Abstract. The U.S. Midwest is experiencing growth in both irrigation and subsurface (tile) drainage. Capturing, storing, and reusing tile drain water, a practice called drainage water recycling (DWR), represents a strategy for supporting supplemental irrigation while also reducing nutrient loads in tile-drained landscapes. This article describes the development and testing of an open-source online tool, Evaluating Drainage Water Recycling Decisions (EDWRD), which integrates soil and reservoir water balances for a tile-drained field and estimates potential benefits of DWR systems across multiple reservoir sizes. Irrigation benefits are quantified by applied irrigation and its relation to the irrigation demand, while water quality benefits are quantified by the amount and percentage of tile drain flow captured by the reservoir. Global sensitivity analysis identified input parameters affecting total available water as the most influential factors in estimating outputs. Initial and mid-season crop coefficients, irrigation management, and reservoir seepage rates were also influential. Curve number, fraction of wetted surface during irrigation, crop coefficients for the end of crop growth and frozen soil conditions, and the non-growing season residue amount were identified as low-sensitivity parameters. Results from the sensitivity analysis were used to prioritize and simplify user interaction with the tool. EDWRD represents the first open-source tool capable of evaluating DWR systems and can be used by multiple user groups to estimate the potential irrigation and water quality benefits of this innovative practice. Keywords: Drainage water recycling, Dual crop coefficient, Open-source model, Sensitivity analysis, Subsurface drainage, Supplemental irrigation.

HighlightsA fuzzy rating system was created based on published criteria for subirrigation suitability.Maps showing potential suitability for subirrigation were created for the U.S. Midwest.78,500 km2 across the U.S. Midwest is potentially suitability for subirrigation.Maps identify potential subirrigation locations pending onsite assessment.Abstract. Subirrigation through subsurface tile drains has potential to increase crop yields and improve water quality in tile-drained landscapes, but it has not been widely implemented. Identifying locations with high potential suitability for subirrigation may help the planning and implementation of this practice. In this study, we developed a fuzzy rating system for subirrigation suitability using the Gridded Soil Survey Geographic Database (gSSURGO). Maps of the fuzzy rating system identified locations of high potential suitability for subirrigation and highlighted physiographic regions highly conducive to the practice. We identified 78,500 km2, about 9%, of agricultural land in the Midwest with high potential suitability for subirrigation where onsite investigation may be targeted. The largest areas of high potential suitability were found in Minnesota, Illinois, and Indiana. Results from the fuzzy rating analysis are provided to the public through three channels: a downloadable data repository, map service, and web map tool. Ultimately, this study can facilitate the adoption of subirrigation by highlighting areas where subirrigation may potentially be a viable practice. Keywords: Controlled drainage, Fuzzy rating, Geographic information system (GIS), Gridded Soil Survey Geographic Database (gSSURGO), Midwestern U.S., Subirrigation.

One of the strategies proposed for reducing nitrate losses from subsurface tile drain systems in Midwestern agricultural lands, controlled drainage, is known to reduce drain flow on an annual basis, but is not well understood for individual events. Understanding hydrologic controls that regulate outflow from free and controlled drainage systems during drainage events can offer improved insight on the overall functioning of these systems so that they can be better managed or retrofitted to increase their environmental benefits. This study used data from a monitoring site in east central Indiana to investigate the hydrology of 22 drainage events in free and controlled subsurface drainage systems. Relationships between event drainage volume, drain flow hydrograph metrics, column soil moisture, water table depth, and precipitation time series for each event were explored to determine the effect of precipitation characteristics and antecedent conditions on these metrics. Controlled drainage reduced event drainage volume and peak flows by 22% ± 12% and 29% ± 16% respectively, and increased the time to peak of drainage by 98% ± 52%. Higher total precipitation and precipitation time spread increased infiltration, while the average precipitation intensity did not correlate with drainage volumes in either system. Peak flow rate in free draining quadrants were positively affected by higher precipitation and average precipitation intensity, while in managed quadrants, the antecedent soil moisture appeared to be more influential in affecting peak flow rate than precipitation characteristics. Surface runoff potential increased with the increase in average precipitation intensity in all quadrants. Saturation excess ponding or possibly overland flow occurred in events that have a low average precipitation intensity, and a high precipitation time spread. Field observations indicate that saturation excess overland flow was more pronounced in controlled quadrants because water table levels rose higher than the water table levels of their free draining counterpart.

CLImate GENerator (CLIGEN) (v5.3), a stochastic weather generator, is widely used in conjunction with the Water Erosion Prediction Project (WEPP) model for runoff and soil loss predictions. CLIGEN generates daily estimates of weather based on long-term observed weather station data statistics. For the United States, the original CLIGEN database released with WEPP in 1995 was derived using inconsistent periods of climate records through 1992 that could lead to significant variations in runoff and soil loss predictions on spatial and temporal scales. To achieve more reliable estimates of runoff and soil loss, an updated climate database was derived from a consistent 40 years of recent climate records of 1974 to 2013 in the United States. The objectives of this study were to (1) examine the spatial patterns in trends of differences in precipitation and maximum and minimum temperatures between the two databases, and (2) evaluate the impacts on WEPP-predicted mean annual runoff and soil loss, from the original to the updated databases. For runoff and soil loss estimates, WEPP simulations were conducted under a tilled fallow condition for 1,600 CLIGEN locations in the contiguous United States for a slope profile of 22.1 m length, 9% slope gradient, and silt loam soil. Comparison of precipitation and maximum and minimum temperatures between the original and updated databases showed variations in spatial patterns both annually and seasonally. Annual precipitation and minimum temperature generally increased across most of the country while maximum temperature increased in the western half of the United States and parts of the Northeast. Seasonally, increases in precipitation are evident in the Midwest in spring, fall, and winter, the Northwest in spring, and the Southeast in fall. Maximum daily temperature has increased in the western half of United States and parts of the Northeast in the winter, fall, and spring, whereas minimum daily temperature has increased in all seasons across the United States. Changes in WEPP-simulated mean annual runoff and soil loss from the use of the original to the updated CLIGEN database showed increases in runoff and soil loss in most of the United States. However, some stations showed either increases or decreases in runoff and/or soil loss with the updated database primarily because of differences in monthly precipitation and intensity values in the two databases. Understanding the impacts of the use of the updated database on runoff and soil loss from this study will help in making informed decisions for conservation planning and management when utilizing the WEPP erosion model.

Controlled drainage (CD), sometimes called drainage water management, is a practice whereby the drainage system outflow is managed during specific periods to retain more water in the field. Although CD has been shown to reduce downstream nitrate-N (NO3--N) load, seasonal patterns have been less consistent which can potentially impact the effectiveness of conservation practices. The main objective of this study was to assess the regional and seasonal impact of conventional free drainage (FD) and CD on drainage flow and nitrate-N load. Using experimental data from ongoing and historical CD experiments across the Corn Belt and in North Carolina, we evaluated subsurface drain flow, nitrate-N load, and performance of CD systems. Across the data set and regions, there was little difference in annual flow from FD conditions. Seasonally, more northern and western sites experienced a greater percentage of the annual flow occurring in the spring. There was no nitrate-N concentration reduction with CD. Flow and nitrate-N load reductions with CD did not vary by plant hardiness zone across the region, but the season with the greatest reduction did shift from winter to spring moving north and west in the study area. Absolute flow reductions (in mm) were similar regardless of precipitation category. Consequently, the percent reduction was lower as the amount of precipitation (category) increased. Overall, this analysis found CD to be an effective practice for reducing drain flow and nitrate-N loading directly delivered by the drains to downstream water bodies across the region.

Subirrigation is employed to supply water to crop root zones via subsurface drainage systems, which are typically installed for the purpose of excess soil water removal. Crop yield increases due to subirrigation have been demonstrated in numerous studies, but there is limited information regarding yield under future climate conditions when growing season conditions are expected to be drier in the U.S. Corn Belt. DRAINMOD was calibrated and validated for three locations with different soil series in northwest Ohio and used to investigate maize relative yield differences between subirrigation and free subsurface drainage for historic (1984–2013) and future (2041–2070) climate conditions. For historic conditions, the mean maize relative yield increased by 27% with subirrigation on the Nappanee loam soil, but had minimal effect on the Paulding clay and Hoytville silty clay soils. Maize relative yield under free subsurface drainage is predicted to decrease in the future, causing the relative yield difference between free subsurface drainage and subirrigation practices to nearly double from 9% to 16% between the historic and future periods. Consequently, the subirrigation practice can potentially mitigate adverse future climate change impacts on maize yield in northwest Ohio.

Accurate quantification of forest aboveground biomass (AGB) is the foundation to the responses of diverse forest ecosystems to the changing climate. Lidar-based statistical models have been used to accurately estimate AGB in large spatial extents, especially in boreal and temperate softwood forest ecosystems. However, the few available models for temperate hardwood and hardwood-dominated mixed forests are low in accuracy due both to the deliquescent growth form of hardwood trees and the strong site-to-site variations in height-diameter relationship. In this study, we established multiplicative nonlinear regression models that incorporated both lidar-derived metrics and soil-based site productivity classes (high and low productivity sites) to estimate aboveground biomass in temperate hardwood forests. The final optimized model had high accuracy (R 2 = 0.81; RMSE = 45.5 Mg ha − 1 ) with reliable performance in ABG estimation by integrating relative height metrics at 75 and 70 percentiles (RH75 and RH70), canopy coverage and site productivity class. An optimized model that included an index of site productivity explained 14% more variance than the best-fit model without the term. Moreover, the relationship between AGB and lidar-based metrics was nonlinear on low productivity sites and nearly linear on high productivity sites, further indicating the importance of including direct or indirect measures of site productivity in lidar-based biomass models, particularly for those applied to temperate hardwood forests. Our new lidar-based model provides a potential framework to integrate lidar-based structural information and soil-based site productivity to improve AGB estimation in temperate hardwood forests.

A two-stage ditch involves modifications of a conventional, trapezoidal drainage ditch to better replicate the features of a natural stream through the addition of adjacent floodplains or benches. Previous research in Indiana and Ohio has shown that two-stage ditches offer the potential to reduce sediment load and extend the interaction time between water, bench vegetation, and bench soil allowing larger uptake of nutrients by the vegetation and increasing the denitrification rates. A two-stage ditch was constructed that drains an area of approximately 267 ha of farmland used for corn and soybean production. Discharge, nitrate-N (NO3), total phosphorus (TP), soluble reactive phosphorus (SRP) and total suspended sediment (TSS) were monitored in the two-stage ditch and a control reach immediately upstream. The two-stage ditch was found to significantly decrease TP, SRP and TSS concentrations and loads. Although the two-stage ditch decreased NO 3 concentrations significantly, it did not have a significant impact on NO3 loads. More specifically, the two-stage ditch reduced the loads of TP by 40%, SRP by 11% and TSS by 22–40% depending on the stage of vegetation establishment on its floodplain benches, compared to an increase in load of 78%, 2% and 1%, respectively in the control reach.

Climate change projections for the Midwest U.S. indicate a future with increased growing season dryness that will adversely impact crop production sustainability. Systems that capture water for later subirrigation use have potential as a climate adaptation strategy to mitigate this increased crop water stress. Three such systems were operated in northwest Ohio from 1996 to 2008, and they exhibited substantial crop yield benefits, especially in dry growing seasons, but also to a lesser extent in near normal or wet growing seasons. The goal of this research was to estimate the increase in crop yield benefits of water capture and subirrigation systems that can be expected under projcted 2041–2070 climate conditions in northwest Ohio. Historical subirrigated field crop yield differences with fields having free drainage only, relative to growing season dryness/wetness, were used to determine future northwest Ohio subirrigated field crop yield increases, based on the modeled climate for 2041–2070. Climate records for 2041–2070 were projected using three bias corrected model combinations, CRCM+CGCM3, RCM3+GFDL, and MM5I+HadCM3. Growing season dryness/wetness was classified based on the difference between rainfall and the crop adjusted potential evapotranspiration using the 1984–2013 climate record at the three system locations. Projected 2041–2070 growing season precipitation varied substantially between the three model combinations; however, all three indicated increased growing season dryness due to rising temperature and solar radiation. The overall subirrigated field corn yield increase rose to an estimated 27.5%–30.0% in 2041–2070 from 20.5% in 1996–2008, while the subirrigated field soybean yield increase improved from 12.2% in 1996–2008 to 19.8%–21.5% for 2041–2070. Consequently, as growing season drought becomes more frequent, the crop yield benefits with water capture and subirrigation systems will improve, and these systems therefore provide a viable climate adaptation strategy for agricultural production.

Conservation practices are designed and implemented to reduce soil erosion and protect water quality, but their effectiveness has been found to vary widely. This study investigated spatial and temporal variability in sediment and nutrient loads and load reductions achieved by implementing conservation practices to understand the implications of landscape heterogeneity and seasonal land-phase hydrologic variation on the effectiveness of the practices. Field-scale effects of nine conservation practices were evaluated in an agricultural watershed in the Midwest using the SWAT model. Results show that the conservation practice effectiveness, and the variability of the practice effectiveness at different locations, vary widely. Most practices are more effective in reducing nutrients in particulate than soluble forms. Practices applied to fields with higher nutrient loads do not necessarily lead to greater load reduction, indicating that an assumption of a proportional relationship between load and expected load reduction on which common targeting strategies for conservation practice placement are based may not be valid for nutrient, in particular soluble forms. The variation of the practices by fields and seasons suggests that achievement of water quality improvement requires careful selection of conservation practices and target areas considering hydrologic variations.

Controlled drainage is a best management practice that decreases nitrate loads from subsurface drainage, but questions remain about optimal operation strategies. One unanswered question is whether the outlet should be lowered prior to or directly after a rainfall event to reduce the amount of time that the water table is at a level that would be detrimental to either trafficability or crop yield. The objective of this study was to determine how much controlled drainage lengthens the time needed for the water table to fall after a rainfall event, to inform possible improvement in the management of controlled drainage systems. This objective was addressed using water table recession rates from two pairs of controlled and free-draining fields located at the Davis Purdue Agricultural Center in Indiana over a period of nine years (2006-2014). At each pair, comparison of mean recession rates from the two fields indicated that controlled drainage reduced recession rate. The significance of the relationship between paired observations and the effect of controlled drainage was determined by a paired watershed approach using analysis of variance (ANOVA) and covariance (ANCOVA). Raising the outlet of the subsurface drainage system decreased the mean rate of water table recession by 29% to 62%, increasing the time needed for the water table level to fall from the surface to 30 and 60 cm depths by approximately 12 to 26 h and 24 to 53 h, respectively. Based on these results, it can be concluded that lowering the outlet before storm events would reduce the amount of time that the water table is at a detrimental level for either crop growth or trafficability. However, the trade-off between costs and benefits of active management depends on the sensitivity of the crop and probability of a severe storm. Keywords: Drainage water management, Managed drainage, Paired watershed approach, Tile drainage, Water table drawdown.

Viable large-scale crop production in the United States requires artificial drainage in humid and poorly drained agricultural regions. Excess water removal is generally achieved by installing tile drains that export water to open ditches that eventually flow into streams. Drainage water management (DWM) is a conservation practice that allows farmers to artificially raise the outlet elevation of a field’s drain tile and can reduce nutrient loss during wet periods by storing more water in the field. We intended to assess the effectiveness of DWM to reduce drainage discharge and nutrient loads and additionally identify predictor variables that influence DWM effectiveness. We compared managed (i.e., DWM) and free draining records using paired t-tests, and identified factors associated with DWM effectiveness using a multiple linear regression approach. T-test results indicated that DWM was highly effective in reducing drainage water discharge and nutrient losses via drain tiles as tile discharge volumes were reduced on average 46%, while tile nitrate loads were reduced by 48%. In addition, total phosphorus and dissolved reactive phosphorus loads were reduced by 55% and 57%, respectively. Based on regression results, we found that several aspects of farm and tile drain management were associated with DWM effectiveness, while site specific landscape characteristics were less likely to predict effectiveness. While DWM is effective as a conservation practice to reduce discharges of water and nutrients from drain tiles, we also identified several knowledge gaps. Future research should investigate effects of DWM on water and nutrients lost in other pathways such as surface runoff, preferential flow, groundwater recharge and biological uptake, and also focus more attention on phosphorus as there is a paucity of research on this topic.

The objective of this study was to evaluate effects of conservation practices actually implemented in reducing sediment and nutrient loads at field and watershed scales. To contribute to the USDA Agency Priority Goal for Water Pilot Projects, we obtained information on conservation practices implemented in the St. Joseph River watershed. Considering expected water quality impacts and simulation ability of the Soil and Water Assessment Tool (SWAT) model, 5,583 of them were selected and incorporated into modeling at a hydrologic response unit (HRU) level by adjusting associated parameters. A calibrated SWAT model was used to estimate load reduction effectiveness of the selected practices. Model results indicated that many of the practices reduced pollutant loads between 10% and 50% at the field scale, with high variability among the practices. Most conservation practices reduced less than 1% of the loads when calculated for the entire watershed, but the load reduction was still large and thus their cumulative long-term effects were expected to be significant. Conservation crop rotation and no-till, which were the most widely applied conservation practices in the study watershed, provided the greatest sediment load reduction, while conservation crop rotation and cover crop reduced the greatest amount of nutrients. Conservation crop rotation, cover crop, no-till, and mulch-till sometimes increased loads of soluble nutrients, resulting in the overall decrease in their effectiveness. Comparison of the spatial distributions of the selected conservation practices and simulated pollutant loads showed existing conservation practices were not targeted for areas producing relatively greater loads. The findings of this study demonstrated different effectiveness of conservation practices at the different spatial scales, suggesting application area, field-scale effectiveness, and placement of the practices are equally critical in achieving watershed-scale water quality improvement.

A survey of the literature and in particular an examination of the terminology use in a previous special collection of modeling calibration and validation articles were conducted to arrive at a list of consistent terminology recommended for writing about hydrologic and water quality model calibration and validation. The terminology list includes rudimentary terms necessary for proper understanding of modeling literature for the novice modeler. This article also provides discussions regarding confusing or conflicting terminology found in the literature, alternative terms to those recommended herein, and alternative definitions for those terms that may be used by some authors.

The Soil and Water Assessment Tool (SWAT) uses hydrologic response units (HRUs) as the basic unit of all model calculations. ArcSWAT, the ArcGIS interface for SWAT, allows users to specify thresholds of land cover, soil, and slope in defining HRUs to improve the computational efficiency of simulations while keeping key landscape features of a watershed in the hydrologic modeling. However, this study found that applying commonly used thresholds in defining HRUs may lead to considerable loss of information about the watershed landscape, emphasizing larger soil types on smaller land covers once the land covers meet a threshold for land cover, and potentially changing average slopes. These changes often have a minor effect on water yield and streamflow simulations by SWAT but a larger effect on sediment and nutrient load simulations, which are more sensitive to slope and soil type and are more influential on outputs at the subwatershed than at the watershed outlet. Study results can help modelers improve their understanding of the HRU strategy for simplifying watershed representation while maintaining major landscape features and make decisions in the HRU delineation process to minimize the chance of biased simulations.

The important questions about agriculture, climate, and sustainability have become increasingly complex and require a coordinated, multifaceted approach for developing new knowledge and understanding. A multistate, transdisciplinary project was begun in 2011 to study the potential for both mitigation and adaptation of corn-based cropping systems to climate variations. The team is measuring the baseline as well as change of the system's carbon (C), nitrogen (N), and water footprints, crop productivity, and pest pressure in response to existing and novel production practices. Nine states and 11 institutions are participating in the project, necessitating a well thought out approach to coordinating field data collection pro- cedures at 35 research sites. In addition, the collected data must be brought together in a way that can be stored and used by persons not originally involved in the data collection, necessi- tating robust procedures for linking metadata with the data and clearly delineated rules for use and publication of data from the overall project. In order to improve the ability to compare data across sites and begin to make inferences about soil and cropping system responses to cli- mate across the region, detailed research protocols were developed to standardize the types of measurements taken and the specific details such as depth, time, method, numbers of samples, and minimum data set required from each site. This process required significant time, debate, and commitment of all the investigators involved with field data collection and was also informed by the data needed to run the simulation models and life cycle analyses. Although individual research teams are collecting additional measurements beyond those stated in the standardized protocols, the written protocols are used by the team for the base measurements to be compared across the region. A centralized database was constructed to meet the needs of current researchers on this project as well as for future use for data synthesis and modeling for agricultural, ecosystem, and climate sciences.

Drainage water recycling, the practice of capturing and storing water drained from fields and using the stored water to irrigate crops when there is a soil water deficit, has been proposed to increase the resiliency of drained agriculture, but the potential benefits have not been quantified. This study determined irrigation and nutrient reduction benefits of drainage water recycling for various reservoir sizes at two tile-drained sites in the U.S. Midwest with differing climates and soils. Field and reservoir water budgets were developed using ten years of measured tile drain flow and weather data. The calculated volume of drain flow that could be captured by the reservoir was combined with measured nitrate-nitrogen and soluble reactive phosphorus concentrations to determine nutrient load reductions. At the Indiana site, a reservoir size representing 6% of the field area (3.05 m depth) would provide water storage for meeting irrigation requirements in all ten years. This reservoir would capture 37% of annual tile drain flow on average, resulting in average annual load reductions of 11 kg ha−1 yr−1 (37%) for nitrate-N and 0.05 kg ha−1 yr−1 (39%) for soluble reactive phosphorus. At the Iowa site, a reservoir size of 8% was necessary to meet the irrigation requirements, which were zero in most years but were higher than at the Indiana site for the three years in which irrigation was needed. This larger reservoir would capture 23% of annual tile drain flow on average, with average annual load reductions of 9 kg ha−1 yr−1 (24%) for nitrate-nitrogen and 0.02 kg ha−1 yr−1 (21%) for soluble reactive phosphorus. Quantifying nutrient load reductions and irrigation potential at these two sites showed that drainage water recycling is a promising practice for the tile-drained landscape of the U.S. Midwest, providing a strategy to manage water-related risks while also contributing to water quality goals.

Drainage water management is a conservation practice that has the potential to reduce drainage outflow and nitrate (NO3) loss from agricultural fields while maintaining or improving crop yields. The goal of this study was to quantify the impact of drainage water management on drain flow, NO3 concentration, and NO3 load from subsurface drainage on two farms in Indiana. Paired field studies were conducted following the paired watershed statistical approach modified to accommodate autocorrelation. Annual NO3 load reductions ranged from 15% to 31%, with an overall reduction of 18% to 23% over the 2-year period, resulting from reductions in both flow and NO3 concentration. Although the study revealed weaknesses in using the paired statistical approach for a dynamic practice like drainage water management, the results of this study support the use of drainage water management as a conservation practice and provide information for decision-makers about the level of benefits that can be anticipated.

Drainage water management, also known as controlled drainage, is the practice of using a water table control structure at the end of the subsurface drain pipe to reduce subsurface drainage, and thereby nitrate losses. Methods to quantify the potential effects of drainage water management for entire watersheds are needed to evaluate the impacts of large-scale adoption. A distributed modeling approach was developed to apply the field-scale DRAINMOD model at the watershed scale, and used to assess the impact of drainage water management on nitrate load from an intensively subsurface drained agricultural watershed in west central Indiana. The watershed was divided into 6460 grid cells for which drain spacing, soil parent material, and cropping pattern were estimated, resulting in 600 unique field conditions. The annual edge-of-field nitrate load from each grid cell was estimated as the product of DRAINMOD-predicted drain flow and the average annual nitrate concentration in drain flow, estimated from observations from related drainage sites in northern Indiana. Predicted monthly streamflow was in good agreement with the observed streamflow (Nash–Sutcliffe efficiency of 0.87 and 0.84 during the calibration and validation periods, respectively) and the predicted drain flow matched well with the measured drain flow (77.1 cm vs. 77.8 cm and 121.3 cm vs. 128.4 cm). Drainage water management decreased the average annual (1985–2009) predicted drain flow from 11.0 to 5.9 cm, and the total nitrate load through subsurface drainage from 236 to 126 ton (both about 47% reduction). The percent reduction in nitrate load varied between 40% and 53% for all combinations of drain spacing, soil parent material and cropping patterns, with drain spacing and soil parent material having a greater effect than cropping pattern. The methodology developed in this study showed potential for predicting the watershed-scale effects of subsurface drainage and drainage water management in drained agricultural watersheds.

The effects of climate variability, drain spacing, and growing season operational strategy on annual drain flow and crop yield were studied for a hypothetical drainage water management (DWM) system at Purdue University9s Water Quality Field Station using the DRAINMOD model. Drainage water management showed potential for reducing annual average (1915–2006) drain flow from all drain spacings (10–35 m) regardless of the growing season operational strategy, with reductions varying between 52 and 55% for the drain spacings considered. Approximately 81 to 99% of the annual drain flow reduction occurred during the non-growing season, depending on the operational strategy. Fixed DWM operational strategies led to an increase in mean predicted yield for narrower spacings compared with conventional drainage systems. Maximum yield was achieved with no control for drain spacings wider than 20 m in 50% of the years. Overall, the height of control had more influence on relative yield than the date of initiation of control. The greatest positive impacts of DWM on relative yield (1.2%) occurred in cool, dry years, while the greatest average negative impacts (−0.2%) occurred in cool, wet years. On average, with the best-case operation selected for annual weather conditions, DWM increased relative yield by approximately 0.8, 0.4, and 0.2% for the 10-, 20-, and 30-m drain spacing, respectively. Accumulated growing degree days and antecedent precipitation index show promise for identifying appropriate operational strategies for DWM.

The hypothetical effects of drainage water management operational strategy on hydrology and crop yield at the Purdue University Water Quality Field Station (WQFS) were simulated using DRAINMOD, a field-scale hydrologic model. The WQFS has forty-eight cropping system treatment plots with 10 m drain spacing. Drain flow observations from a subset of the treatment plots with continuous corn (Zea mays L.) were used to calibrate the model, which was then used to develop an operational strategy for drainage water management. The chosen dates of raising and lowering the outlet during the crop period were 10 and 85 days after planting, respectively, with a control height of 50 cm above the drain (40 cm from the surface). The potential effects of this operational strategy on hydrology and corn yield were simulated over a period of 15 years from 1991 to 2005. On average, the predicted annual drain flows were reduced by 60% (statistically significant at 95% level). This is the most significant benefit of drainage water management since it may reduce the nitrate load to the receiving streams. About 68% of the reduced drain flow contributed to an increase in seepage. Drainage water management increased the average surface runoff by about 85% and slightly decreased the relative yield of corn crop by 0.5% (both are not statistically significant at 95% level). On average, the relative yield due to wet stress (RYw) decreased by 1.3% while relative yield due to dry stress (RYd) increased by 1%. Overall, the relative crop yield increased in 5 years (within a range of 0.8�6.9%), decreased in 8 years (within a range of 0.2�5.5%), and was not affected in the remaining 2 years. With simulated drainage water management, the water table rose above the conventional drainage level during both the winter and the crop periods in all years (except 2002 crop season). The annual maximum winter period rise ranged between 47 cm (1995) and 87 cm (1992), and the annual maximum crop period rise ranged between no effect (2002) and 47 cm (1993).

DRAINMOD was run for 15 years to predict and compare drain flow for three drain spacings and crop yield for four drain spacings at the Southeastern Purdue Agricultural Center (SEPAC). Data from two continuous years of daily drain flow from one spacing were used to calibrate the eight most uncertain parameters using a multi-objective calibration function and an automatic calibration method. The model was tested using the remaining field data for the 5, 10, and 20 m drain spacings for drain flow and the additional 40 m spacing for yield predictions. Nash–Sutcliffe efficiency (EF) for daily drain flow simulations for the calibration years and drain spacing ranged from 0.62 to 0.79. The daily EF for model testing ranged from −0.66 to 0.81, with the average deviations of 0.01 to 0.07 cm/day and standard errors of 0.03–0.17 cm/day. On a monthly basis, 91% of plot years had EF values over 0.5 and 76% over 0.6 for years with on-site rainfall data. The total yearly drain flow was predicted within ±25% in 71% of plot years, and within ±50% in 93% of plot years with on-site rainfall data. Statistical tests of daily drain flow EF values for three spacings and percent errors of crop relative yield for four spacings indicated that the reliability of the model is not significantly different among different spacings, supporting the use of DRAINMOD to study the efficiencies of different drain spacings and to guide the drain spacing design for specific soils. In general, the model correctly predicted the pattern of yearly relative yield change. The relative corn ( Zea mays L.) and soybean ( Glycine max L.) yields were well predicted on average, with percent errors ranging from 1.3 to 9.7% for corn and from −3.3 to 10.3% for soybean.

Structural best management practices (BMPs) that reduce soil erosion and nutrient losses have been recommended and installed on agricultural land for years. A structural BMP is expected to be fully functional only for a limited period after installation, after which degradation of the BMP is likely to lead to a reduction in the water quality improvement provided by the BMP. Assessing the impact of BMPs on water quality is of widespread interest, but no standard methods exist to determine the water quality impact of structural BMPs, particularly as the impact changes through time. The objective of this study was to determine the long-term (~20 year) impact of structural BMPs in two subwatersheds of Black Creek on sediment and phosphorus loads using the Soil and Water Assessment Tool (SWAT) model. The BMPs were represented by modifying SWAT parameters to reflect the impact the practice has on the processes simulated within SWAT, both when practices are fully functional and as their condition deteriorates. The current condition of the BMPs was determined using field evaluation results from a previously developed BMP condition evaluation tool. Based on simulations in the two subwatersheds, BMPs in good condition reduced the average annual sediment yield by 16% to 32% and the average annual phosphorus yield by 10% to 24%. BMPs in their current condition reduced sediment yield by only 7% to 10% and phosphorus yield by 7% to 17%.

Although the U.S. Congress established the Total Maximum Daily Load (TMDL) program in the original Clean Water Act of 1972, Section 303(d), it did not receive attention until the 1990s. Currently, two methods are available for tracking pollution in the environment and assessing the effectiveness of the TMDL process on improving the quality of impaired water bodies: field monitoring and mathematical/computer modeling. Field monitoring may be the most appropriate method, but its use is limited due to high costs and extreme spatial and temporal ecosystem variability. Mathematical models provide an alternative to field monitoring that can potentially save time, reduce cost, and minimize the need for testing management alternatives. However, the uncertainty of the model results is a major concern. Uncertainty is defined as the estimated amount by which an observed or calculated value may depart from the true value, and it has important policy, regulatory, and management implications. The source and magnitude of uncertainty and its impact on TMDL assessment has not been studied in depth. This article describes the collective experience of scientists and engineers in the assessment of uncertainty associated with TMDL models. It reviews sources of uncertainty (e.g., input variability, model algorithms, model calibration data, and scale), methods of uncertainty evaluation (e.g., first-order approximation, mean value first-order reliability method, Monte Carlo, Latin hypercube sampling with constrained Monte Carlo, and generalized likelihood uncertainty estimation), and strategies for communicating uncertainty in TMDL models to users. Four case studies are presented to highlight uncertainty quantification in TMDL models. Results indicate that uncertainty in TMDL models is a real issue and should be taken into consideration not only during the TMDL assessment phase, but also in the design of BMPs during the TMDL implementation phase. First-order error (FOE) analysis and Monte Carlo simulation (MCS) or any modified versions of these two basic methods may be used to assess uncertainty. This collective study concludes that a more scientific method to account for uncertainty would be to develop uncertainty probability distribution functions and transfer such uncertainties to TMDL load allocation through the margin of safety component, which is selected arbitrarily at the present time. It is proposed that explicit quantification of uncertainty be made an integral part of the TMDL process. This will benefit private industry, the scientific community, regulatory agencies, and action agencies involved with TMDL development and implementation.

A two-step global sensitivity analysis was conducted for the nitrogen simulation model DRAINMOD-N II to assess the sensitivity of model predictions of NO3-N losses with drainage water to various model inputs. Factors screening using the LH-OAT (Latin hypercube sampling - one at a time) sensitivity analysis method was performed as a first step considering 48 model parameters; then a variance-based sensitivity analysis was conducted for 20 model parameters, which were the parameters ranked 1 to 14 by the LH-OAT method, five organic carbon (OC) decomposition parameters, and the empirical shape factor of the temperature response function for the nitrification process. DRAINMOD-N II simulated a continuous corn production on a subsurface drained sandy loam soil using a 40-year (1951-1990) eastern North Carolina climatological record. Results from the first 20-year period of the simulations were used to initialize the soil organic matter pools, and results from the last 20-year period of the simulations were considered for the sensitivity analyses. Both yearly and 20-year average model predictions of NO3-N losses through drainage flow were used in the analyses. Both sensitivity analysis methods indicated that DRAINMOD-N II is most sensitive to denitrification parameters, especially those controlling temperature effect on process rate. Results also indicated that the model is mildly sensitive to the parameters controlling OC decomposition and associated N mineralization/immobilization. The use of different sensitivity analysis methods with dissimilar theoretical foundations increases the confidence in key parameters identification. More efforts should be focused on quantifying key parameters for more accurate model predictions.

Backpropagation neural networks with five data coding schemes were used to predict maize yield at three scales in east-central Indiana of the Midwest USA, using 1901–1996 local crop-stage weather data and yield data from farm, county, and state levels. Input data included precipitation and air temperature during maize reproductive (R) stages R1 (silking) to R5 (denting of kernels), the year, and, for some nets, the scale of yield data. The five coding schemes were maximum value, maximum and minimum value, logarithm, thermometer (powers of 10), and binary (powers of 2). Root mean squared error over a testing set was determined at farm, county, and state scales. The best version of the network was maximum and minimum value coded and gave a root mean squared error of 10·5% overall (8·6% farm, 12·5% county, 9·0% state yield). The prediction error among the five coding types ranged from 10·5 to 46·9% for the best net of each type. Neural net software usually has a default coding scheme, which is used without considering an alternative. The results of this study suggested that the data coding method had a significant effect on neural net performance, and that sensitivity testing of data representation should be performed when constructing neural nets. The study also confirmed the usefulness of neural nets for yield prediction from simple data sets.

The Soil Moisture Routing Model (SMRM) and a modified SMRM that included the effect of clay outcrop were used to model runoff source areas for watersheds W103 and W110 located within the North Appalachian Experimental Facility (NAEW) in Coshocton, OH, USA. The two adjacent watersheds have visually similar landscape and soil types but exhibit very different runoff responses. Daily measured and predicted runoff for nineteen years (1980-1998) and periodic measured and estimated soil moisture at three locations in each watershed for four years (1975-1978) were compared using Nash-Sutcliffe efficiency (R 2 ) to evaluate adequacy of model prediction. Overall, the performance of the model was reasonable. Daily runoff was predicted better at W103 (R 2 = 0.59) than at W110 (R 2 =0.47). Soil moisture at all locations was also reasonably predicted (R 2 > 0.5). Most of the high runoff generating areas in W103 and W110 were located in the middle and lower landscape positions.