|
Drainage Water Management: A Review of Nutrient Load Reductions and Cost Effectiveness
Jane Frankenberger1,*, Sara K. W. McMillan2, M. R. Williams3, Katy Mazer1, Jared Ross4, Brent Sohngen5
Published in Journal of the ASABE 67(4): 1077-1092 (doi: 10.13031/ja.15549). 2024 American Society of Agricultural and Biological Engineers.
1 Agricultural and Biological Engineering, Purdue University, West Lafayette, Indiana, USA.
2 Department of Agricultural and Biosystems Engineering, Iowa State University, Ames, Iowa, USA.
3 National Soil Erosion Research Laboratory, USDA ARS, West Lafayette, Indiana, USA.
4 Fisheries and Wildlife, Michigan State University, East Lansing, Michigan, USA.
5 Agricultural, Environmental, and Development Economics, Ohio State University, Columbus, Ohio, USA.
* Correspondence: frankenb@purdue.edu
The authors have paid for open access for this article. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License https://creative commons.org/licenses/by-nc-nd/4.0/
Submitted for review on 21 January 2023 as manuscript number NRES 15549; approved for publication as a Review Article and as part of the Agricultural Conservation Practice Effectiveness Collection by Community Editor Dr. Ruth Book of the Natural Resources & Environmental Systems Community of ASABE on 28 June 2023.
Mention of company or trade names is for description only and does not imply endorsement by the USDA. The USDA is an equal opportunity provider and employer. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the USDA.
Highlights
Abstract.Drainage water management (DWM) is the practice of seasonally adjusting the drainage outlet elevation with a control structure and is implemented to decrease nutrient loss in drained agricultural landscapes. In this study, we review the effect of DWM on annual nitrogen (N) and phosphorus (P) loads compared to free drainage and examine its cost-effectiveness using data from around the world published between 1979 and 2022. Studies included in the review were limited to row-crop agriculture without irrigation, with controlled drainage and free drainage compared during the same year. A total of 290 plot-years for N were compiled and analyzed using a mixed effects linear model to assess differences in flow, load, and concentrations and account for within and among site variability. Few studies examined dissolved reactive P (n=7) or total P (n=3); thus, only DRP results from these studies were statistically evaluated using Kruskal-Wallis tests. Analysis of N data showed that DWM reduced N loss on average by 13.3 kg/ha/yr (95% confidence interval: 9.4-17.3 kg/ha/yr), which corresponds to a 46% load removal efficiency (95% confidence interval: 37.4%-54.5%). Drainage water management decreased annual DRP and TP load compared to free drainage by an average of 0.04 kg/ha/yr (mean=34%; range=-35% to 80%) and 0.06 kg/ha/yr (mean=36%; range=7% to 72%), respectively. Economic analysis showed that the cost of N reduction would range from $1 to $11 per kg N removed assuming no yield benefit, while a 3% yield increase would lead to a likely net economic benefit to the producer. Findings indicate that DWM can be an effective practice for decreasing nutrient loading from drained landscapes, however, losses through unmeasured flow paths such as seepage and surface runoff may lead to overestimating the effectiveness of DWM. Additionally, complex interactions among soil organic matter, soil nutrient levels, and water table depths need to be accounted for to ensure both N and P removal is achieved. Recommendations for monitoring discharge and nutrient loads in future DWM studies are provided to better understand the processes influencing nutrient load reductions and/or increases and further quantify the effectiveness of DWM implementation in drained landscapes.
To address nitrogen (N) and phosphorus (P) losses from agricultural lands that cause eutrophication and algal blooms, conservation practices are often implemented both in the field and at the edge-of-field (Feyereisen et al., 2022). In this study, we focus on drainage water management (DWM), a practice that is also known as controlled drainage, and its effect on decreasing N and P loads from non-irrigated row crop agricultural fields to receiving surface waters. DWM aims to decrease nutrient loads exported from fields by modifying the management of artificial drainage systems that are installed in poorly drained soils to support trafficability and crop growth. Conventional drainage systems can be a significant source of both N (e.g., Williams et al. [2015a]) and P losses (e.g., King et al. [2015]). This article is part of a collection that provides a systematic review and evaluation of the performance and cost effectiveness of agricultural conservation practices on nutrient and sediment reduction.
DWM is the process of controlling the outlet elevation of a surface or subsurface agricultural drainage system to manage the drainage volume and water table elevation. It can be implemented using USDA NRCS Conservation Practice Standard 554 (USDA-NRCS, 2020). It is typically applied by installing a structure or series of structures in a drainage ditch or subsurface tile drain, allowing the drainage outlet elevation to be adjusted or managed (fig. 1). The premise behind this practice is that the same drainage intensity is not required at all times (Skaggs et al., 2012) and that it is possible to decrease drainage rates during some parts of the year without negatively affecting crop production. The main goal of DWM is therefore to improve water quality in downstream waters by raising the outlet in the non-growing season or after the crop is planted to decrease drainage rates and potentially conserve water for crop use during the growing season. DWM is applicable to agricultural lands with surface or subsurface agricultural drainage systems that can be adapted to allow for management of the outlet elevation (Frankenberger et al., 2023). Structures are typically needed for every 30 to 60 cm change in elevation. Thus, DWM implementation is suitable for areas where a high natural water table exists (or existed) and the topography is relatively smooth and flat to very gently sloping (i.e., <1%; unless drainage laterals are installed on the contour).
 |
| Figure 1. Field with drainage water management in a subsurface drainage system, showing water control structures raising the outlet and water table (image from the Transforming Drainage project). |
DWM was originally called controlled drainage in the 1940s and 1950s, when it was primarily used to control subsidence in peat and muck soils. Research in Florida (Clayton and Jones, 1941), Indiana (Jongedyk et al., 1954), and other locations (Stephens, 1955) showed its effectiveness for limiting subsidence, with Willardson et al. (1972) showing that denitrification was taking place in what they called submerged drains. Research first showing that DWM could be effective for decreasing nutrient loss in drainage water on non-irrigated land was conducted in North Carolina (Gilliam et al., 1979) on surface ditches. Research on this practice has since expanded for use on subsurface tile drains and to other poorly drained regions of North America and Europe.
Over the years, several different terms have been used for DWM, resulting in some confusion. For example, the term controlled drainage has been used to describe a related but different practice that includes subirrigation – i.e., controlled drainage and subirrigation (CD-SI), which combines the control of the elevation of the drainage outlet with adding water to irrigate the crop through the same drains (e.g., Satchithanantham et al. [2014]). However, the inclusion of subirrigation makes DWM and CD-SI quite different. The terms water table management (WTM), controlled water table (CWT), or controlled and reversible drainage (CaRD; Doty, 1980) were also used throughout the 1980s and 1990s, usually referring to CD-SI. Starting in 2002, researchers, implementation agencies, and the drainage industry coordinated efforts through the Agricultural Drainage Management Systems Task Force to increase use of controlled drainage (without subirrigation) in the U.S. Midwest to decrease N delivery to the Gulf of Mexico (Fouss and Sullivan, 2009). The focus was on water quality and conservation through controlled drainage alone, in contrast to the earlier efforts focusing on increasing crop yields using CD-SI, which requires a water source and has a more limited potential application. The name drainage water management was used in the NRCS Conservation Practice Standard 554 developed during this period to be consistent with other NRCS conservation practices like irrigation water management and to alleviate concerns producers might have about the word control. Thus, many published studies began using the term drainage water management, while others continued to use controlled drainage orcontrolled tile drainage (CTD) (e.g., Sunohara et al. [2014]). In this article, the term drainage water management (DWM) refers to controlled drainage without subirrigation.
Several reviews on the effectiveness of DWM to decrease nutrient loads have been written since 2012 (Skaggs et al., 2012 [N only]; King et al., 2015 [P only]; Smith et al., 2019; Ross et al., 2016 [N and P]; Carstensen et al., 2020 [N and P]; Wang et al., 2020 [N and P]); Kesicka et al., 2022 [modeling studies only]). Drainage water management studies included in the Transforming Drainage dataset (Chigladze et al., 2021; Abendroth et al., 2022; http://drainagedata.org) have been synthesized by Helmers et al. (2022) for N loss, and Youssef et al. (2023) for crop yield effects. Previous reviews have varied considerably in the papers included, as described in supplemental table S1. For instance, some have included studies of CD-SI, paddy rice, and column or lysimeter studies, which are not included here and confound the interpretation of findings. There have been several recent studies since the previous reviews, and all reviews left out some of the studies included in the current analysis. Data summarized as part of previous reviews have only been provided in publicly-accessible databases by Cartensen et al. (2020) and Smith et al. (2019), and this lack limits future analysis and expansion of the datasets.
This review builds on this previous work to determine the cost-effectiveness of DWM and examine the processes and factors affecting nutrient load reductions. Specific objectives were to (1) compare, integrate, and synthesize results from studies conducted under different experimental conditions and site conditions to obtain a systematic understanding of the mitigation efficacy of DWM; (2) discuss tradeoffs between N and P associated with DWM implementation, (3) determine performance-based costs of DWM; (4) provide monitoring recommendations for future DWM studies to constrain variability and identify environmental drivers that enhance effectiveness; and (5) provide summarized data in a publicly-available dataset that can be expanded and analyzed as new studies are completed. Information gained from this review can be used to develop recommendations for cost-effective conservation practices to be considered for prioritization when funding agencies are developing their programs. This review also provides support for updates to technical content in the USDA NRCS conservation practice standards, and guidance for agencies and organizations by documenting the magnitude of the nutrient pollution reduction due to DWM.
Review Methodology
Search Strategy and Inclusion Criteria
A literature search was conducted using key words including “drainage water management,” “controlled drainage,” “controlled tile drainage,” “water table management,” “water table control,” “controlled agricultural drainage,” or “drainage control,” as well as references from published studies. All studies were located in the eastern and midwestern US, eastern Canada, and Europe, where precipitation allows cultivation without irrigation and drainage is required for crop production. We selected only studies that included measured data rather than simulated data or modeling results. Additional criteria included comparison of free drainage and controlled drainage in the same year with consistent drainage treatments for at least one year, detailed descriptions of the experimental design, and reporting of annual results for N and/or P (total P [TP] or dissolved reactive P [DRP]) loads measured at the drain outlet. This excluded three Canadian studies that were monitored only during the growing season. Subsurface tile drainage was used in most studies, although N studies using surface ditch drainage in Italy and North Carolina were also included. We excluded studies of rice paddies and irrigated cropland where controlled drainage is also used (e.g., Ayars et al. [2006]) because controlled drainage in these settings is expected to have different effects than DWM in rainfed cropland.
Studies that combined subirrigation with controlled drainage (CD-SI) were excluded, which omitted many studies included in previous reviews. Subirrigation can change the effect of the practice through multiple mechanisms, most notably because providing additional water leads to greater and more consistent yields in the treatment plots, thereby affecting both nutrient loads and the water balance. Greater yields increase nutrient uptake, and this effect cannot be distinguished from the effect of the water management treatment alone. For studies that included three treatments (free drainage, DWM, and CD-SI) (Drury et al., 2009; Fausey, 2005), we only included DWM treatment results in this review. Lastly, we excluded studies on land uses other than row crop agriculture, which is most commonly a corn-soybean rotation in the U.S. but included barley, oats, and wheat in Europe. A list of studies included in the current study compared to previous reviews is shown in supplemental table S1.
Search criteria resulted in the identification of 31 sites where DWM and free drainage were compared in side-by-side treatments and documented in papers or datasets published between 1979 and 2022 (fig. 2; table 1). All of these sites reported treatment effects on N loads, and 7 also reported P loads. Because many years of research are needed to generate enough data to understand a practice, several papers may present data from the same research site. In some cases, the data for all years are included in a subsequent paper (e.g., all four years for SW1 are provided in Wesstrom and Messing [2007], although some were published earlier), but in others, the data needed to be accessed from more than one paper for all years (e.g., the nine years of IA2 data are divided between Helmers et al. (2012) (data for 2007-2010) and Schott et al. (2017) (data for 2011-2015). This review and resulting dataset are therefore organized by research site, with the paper(s) listed in a separate column. Combining data by the sites at which the research was conducted rather than by paper will help future users of the data understand the studies and provide a basis for further analysis as new data are generated. After accounting for any overlap among papers from the same site, 290 site-years of data were used for the N analysis.
Data Extraction and Statistical Analysis
Annual losses of N and P were extracted for both the DWM (treatment) and free drainage (control) plots, except for papers that reported cumulative averages for their multi-year study as a single data point (i.e., Fausey, 2005). We compared replicates as the study authors presented in the papers, unless they were not paired, in which case we compared adjacent plots. Annual discharge and nutrient loading can be problematic for interpreting the effects of DWM because management is not consistent throughout the year, as the water held back in the DWM plot must be released before field operations. However, comparing annual values was the only viable comparison given varied management across sites and years and different reporting by authors. Annual values have also been used in previous reviews (Ross et al., 2016; Cartensen et al., 2020; Wang et al., 2020).
Nitrate was the only form of N monitored consistently; both DRP and TP were included where reported. Annual losses were extracted from the papers using tables, when available, or from figures. Often, nutrient concentrations were not reported, so we calculated a flow weighted mean annual concentration using the annual nutrient load and discharge volume. In addition to collecting discharge and water quality response variables, we also included environmental factors that may influence DWM effectiveness, such as soil texture, drainage class (e.g., poorly drained), and cropping system. Drainage intensity, a measure of the rate at which water can move through the soil to the drains that was recommended by Skaggs (2017) for drainage analysis, was calculated using the tool at https://transformingdrainage.org/tools/drainage-rate/. A plot or field area was used as reported in the paper, although the entire area may not have been effectively controlled. Load reductions were calculated as a difference between DWM and free drainage (FD), with positive values indicating a reduction in mass loss. Removal efficiency was calculated as follows:
(1)
 |
| Figure 2. Sites with comparisons of N or P reductions from drainage water management in the United States, Canada, and Europe. Symbol size indicates the number of years of comparison between controlled and free drainage, and shape/color indicates whether both N and P were included or only N. |
| Table 1. Sites included in this review and dataset. 31 studies met all criteria, leading to a total of 290 plot-years included in this synthesis. | ||||||
| Site | Key Reference (additional references in table S1) | Location Name | Param. | Years | Plots Reported Separately[a] | Years Reported |
| DEN1 | Carstensen et al. 2019 | Odder, Denmark | N, P | 2012-2015 | 4 | 3 |
| IA1 | Jaynes, 2012 | Story County, IA | N | 2006-2009 | 2 | 4 |
| IA2 | Helmers et al., 2012, Schott et al., 2017 | SERF (Crawfordsville, IA) | N | 2007-2015 | 2 | 11 |
| IL 1 | Cooke and Verma, 2012 | Barry, IL | N | 2008-2009 | 2 | 2 |
| IL 2 | Cooke and Verma, 2012 | Enfield, IL | N | 2008-2009 | 2 | 2 |
| IL 3 | Cooke and Verma, 2012 | Hume, South IL | N | 2008-2009 | 2 | 2 |
| IL 4 | Cooke and Verma, 2012 | Hume, IL | N | 2008-2009 | 2 | 2 |
| IL5 | Woli et al., 2010 | Deland, Piatt Cty, IL | N | 2007-2009 | 2 | 3 |
| IN1 | Adeuya et al., 2012 | White County, IN - Site A | N | 2007-2009 | 2 | 2 |
| IN2 | Adeuya et al., 2012 | White County, IN - Site B | N | 2007-2009 | 2 | 2 |
| IN3 | Saadat et al., 2018 | DPAC, Randolph Cty, IN | N, P | 2007-2016 | 4 | 11 |
| IT1 | Tolomio and Borin, 2018 | Legnaro, NE Italy | N, P | 2007-2013 | 4 | 5 |
| LI1 | Ramoska et al., 2011 | Middle Lithuania | N | 2001-2007 | 2 | 7 |
| MN1 | Strock et al., 2022 | Redwood Cty, MN | N | 2006-2017 | 2 | 10 |
| MN2 | ADMC, 2012 | Dundas, MN | N | 2008-2009 | 2 | 2 |
| MN3 | ADMC, 2012 | Hayfield, MN | N | 2008-2009 | 2 | 2 |
| MN4 | ADMC, 2012 | Wilmont, MN | N | 2008-2009 | 2 | 2 |
| MN5 | ADMC, 2012 | Windom, MN | N | 2008-2009 | 2 | 1 |
| MO1 | Nash et al., 2015b (N Loss), Nash et al., 2015a (P loss) | Univ of MO Greenley Res. Center, Novelty, MO | N, P | 2010-2013 | 2 | 3 |
| NC1 | Gilliam et al., 1979 | Kinston, NC | N | 1974-1975 | 2 | 1 |
| NC2 | Gilliam et al., 1979 | Plymouth, NC | N | 1973-1976 | 2 | 1 |
| NC3 | Evans et al., 1989 | Cahoon, Pamlico Cty, NC | N | 1987 | 2 | 1 |
| NC4 | Evans et al., 1989 | Reid, Camden Cty, NC | N | 1987 | 2 | 1 |
| NC5 | Poole et al., 2018 | Tidewater Research Station, Plymouth, NC | N | 1992-94; 2007-2012 | 4 | 9 |
| OH1 | Fausey, 2005 | OARDC, Wood Cty, OH | N | 1999-2003 | 2 | 4 |
| OH2 | Williams et al., 2015a | Upper Big Walnut Creek, OH | N, P | 2009-2012 | 4 | 2 |
| ON2 | Tan et al., 1998 | Southwestern Ontario | N | 1995-96 | 2 | 3 |
| ON3 | Drury et al., 2009 | Whelan Experimental Farm, Woodslee, Ontario | N | 1995-99 | 2 | 2 |
| SD1 | Sahani, 2017; TD database | SERF, Clay Cty, SD | N, P | 2014-2016 | 4 | 3 |
| SD2 | Sharma, 2018 | Hanson Cty, SD | N | 2016-17 | 2 | 4 |
| SW1 | Wesstrom and Messing, 2007 | Manstorp, Sweden | N, P | 1996-2000 | 2 | 11 |
[a] 1 plot may contain replicates (e.g., Drury et al. [2009] had 4 replicates of each treatment and Helmers et al. [2012] had 2 replicates of each treatment. | ||||||
Reductions and percent changes were also calculated for discharge volume and nutrient concentration between DWM and free drainage treatments.
The 290 individual plot years resulted in 140 paired plot-years of data for NO3-N load, 127 for concentration, and 131 for flow volume, which allowed us to perform robust statistical analyses on those data. All analyses were performed in R software version 4.0.5, specifically the lme4 package in R (Bates et al., 2015; R core Team). The limited number of papers on the effectiveness of DWM on P loads did not allow us to perform robust statistical analysis. Rather, we qualitatively compare and summarize the results from those papers and provide recommendations to gain a better understanding.
The reductions in flow, load, and concentration are defined as the effect size for this synthesis. Exploratory data analysis included tests of normality, and all the data were positively skewed. To account for this non-normality and to also test for independence (i.e., does correlation across years for a single site need to be accounted for or can each plot-year be treated as an independent observation), we compared the performance of two models: fixed and mixed effects linear models. A linear mixed model estimating reductions for load, concentration, flow, and load removal efficiency was developed with site as a random effect using the lme4 package in R (Bates et al., 2015), and compared to a fixed effects model. A chi-squared test showed that mixed and linear effects models were significantly different for load and flow (p < 0.001), suggesting that the mixed effects model was needed to account for the non-independence for plot years at the same site. Although mixed and fixed effects models were not significantly different for concentration (p = 0.22), a mixed effects model was used as well for consistency. Another assumption of these models is that the variance of the residuals is constant (i.e., homoscedasticity). For each model, we plotted predicted versus measured values and visually inspected these plots. While there was slightly greater variability at higher concentrations, loads, and flow, variance was generally similar. In addition, the number of paired plot-years per site was relatively small, suggesting that within-site variance was relatively constant for all sites. We report the variance in the effect size at the 95% confidence interval around the mean difference between CD and FD. The mean, standard error, and confidence interval were obtained from the mixed effects model.
Data Availability
The dataset used in this analysis is provided in the Supplemental Materials at https://doi.org/10.13031/23596572.v1 and also at https://conservationdrainage.net/resources/controlled-drainage/reviewdata, so other researchers can continue to build on this analysis to better understand the impacts of DWM. Information provided in the dataset includes field or plot size, location (latitude and longitude), study duration, discharge and nutrient load for the treatment and control plot by year, and as many of the following as could be identified: drainage depth and spacing, cropping system, and soil properties (taxonomy, drainage class, texture, sand/silt/clay percentages, hydraulic conductivity).
Nutrient Reduction Effectiveness
Nitrate-Nitrogen
Observations of Paired DWM Plots
Drain flow, NO3-N concentration, and load were compared for all FD and CD paired plot-years. In general, median FD flows were higher than CD flows (263 compared to 165 mm/yr, respectively), and NO3-N loads followed a similar trend (23.4 compared to 11.9 kg N/ha/yr, respectively) (fig. 3). NO3-N concentrations were generally similar between drainage types (9.0 compared to 8.9 mg/L, respectively), although some studies in this synthesis documented changes in concentration. For example, Wesstrom and Messing (2007) observed rising groundwater coupled with decreasing NO3-N concentration in the late spring in controlled drainage plots. This seasonal variability was also observed by Adeuya et al. (2012), with lower concentrations during the dormant season for both CD and FD, which they attributed to higher water tables, anaerobic soil conditions, and NO3-N loss via denitrification.
Paired plot reductions show that flow and load generally have positive reductions (i.e., CD < FD), while concentrations are centered around 0 (fig. 4). Median values were 95 mm/yr, 0.2 mg/L, and 9.9 kg/ha for flow, concentration, and load reductions, respectively. The distribution of percent load reduction was very similar to percent flow, further suggesting that load reductions are driven by flow. This is also supported by the observation that the median concentration difference is near zero.
Bivariate plots of CD versus FD further illustrate the reduction in load and flow as a function of DWM (fig. 5). For visual reference, we have included the linear correlation and 1:1 line for each plot. Concentration values are centered around the 1:1 line, suggesting that there was no effect on concentration from DWM. However, both flow and NO3-N load show a large proportion of points below the 1:1 line, which mean that flow and load are lower in CD plots compared to FD.
Effects of DWM on Load, Flow, and Concentration Reductions
Results from the mixed effect model showed that controlled drainage reduced NO3-N loads consistently across the dataset with a mean load reduction of 13.3 kg/ha/yr (95% confidence interval of [9.4, 17.3] and a p-value < 0.001) and 46% load reduction efficiency (95% confidence interval of [37.4, 54.5] and a p-value < 0.001) (fig. 6). This load reduction value is higher than the median value (9.9 kg/ha/yr) but more robust because the mixed effects model accounts for differences among sites, non-independence within sites, and non-normality of the data. This analysis showed similar results, with load reductions closely mirroring reductions in flow volume and no effect on NO3-N concentrations (figs. 4 and 5). The mixed effects model showed a mean flow reduction of 126 mm/yr, with a 95% confidence interval of (91, 162) and a p-value of <0.001. The grand mean of the mixed effect model for paired concentration reduction was 0.31 mg/L with a 95% confidence interval of (-0.4, 1.1), although the effect was not significant (p = 0.37). These observations are consistent with other recent reviews (Ross et al., 2016; Carstensen et al., 2019) and suggest that there may be additional flow paths (e.g., seepage or surface runoff) that are unaccounted for.
 |
| Figure 3. Boxplots showing distribution of tile flow (mm/yr), NO3-N concentration (mg/L), and NO3-N loads (kg/ha/yr) for free and controlled drainage plot years. Gray lines connect paired values. The solid line in the boxplot shows the median, and the whiskers represent 1.5*interquartile range. |
 |
| Figure 4. Boxplots showing reductions between FD and CD from tile flow (mm/yr), NO3-N concentrations (mg/L), and NO3-N loads (kg/ha/yr) for all paired plot years in the dataset. The solid line in the boxplot shows the median, and the whiskers represent 1.5*interquartile range. |
Although environmental factors were inconsistently reported, we were able to assess the effect of a subset of them on NO3-N load reductions. NO3-N concentration and soil carbon interact with management practices to influence microbial biomass and activity. While none of the studies reported organic matter or soil carbon, we were able to investigate the influence of NO3-N concentrations. Net load reductions were greater at sites with higher concentrations simply because there was more NO3-N entering the system (Spearman’s rho = 0.34, p < 0.001). Drury et al. (2009) investigated the impact of NO3-N fertilization rates coupled with DWM and showed modest reduction in efficiency (31% for higher fertilization rates compared to 44% for lower rates). We hypothesized that efficiency would decrease at higher NO3-N concentrations as substrate supply outpaces denitrification and microbial demand; however, our synthesis showed no correlation between removal efficiency and concentration (fig. 7; Spearman’s rho = 0.071, p = 0.41). Although some studies have noted enhanced denitrification in plots with DWM (e.g., Gilliam et al. [1979] and Carstensen et al. [2019]), the majority do not report lower concentration or higher denitrification rates. This outcome is likely due in part to insufficient amounts of soil C and the absence of anoxic zones in surface soils.
 |
| Figure 5. Bivariate plots of (a) drain flow, (b) NO3-N concentration, and (c) NO3-N load for pair plot-years. Linear correlation (forced through the origin) and the 1:1 line are shown for visual reference. |
 |
| Figure 6. Distribution of NO3-N load reductions in kg N/ha/yr (left) and percent of NO3-N load in free drainage (right). Sites are arranged from highest to lowest median load reduction in kg N/ha/yr. Boxes show 25th and 75th percentile, the whiskers are +/-1.5 x the interquartile range, and all points outside those values are shown. The blue lines show the grand mean and confidence interval from the mixed effect model. |
We analyzed the effect of soil drainage class on NO3-N load reduction in DWM by identifying a representative soil drainage class for sites in the U.S., using soil survey information that was not provided in the paper. This analysis showed that load reductions were slightly higher for less poorly drained soils (mw, swp) and lower for more poorly drained soils (p, vp), though an ANOVA showed that these differences were significant only at the 90% confidence level (fig. 8; p = 0.071). We suggest that the higher reduction in the less poorly drained soils is not due to DWM effectiveness but rather due to seepage losses in these soils, where the restrictive layer is not as impermeable as in very poorly drained soils. Seepage losses may appear in the data as load reductions. The effect of drainage class on percent reduction was not significant (p = 0.245). Drainage intensity was also assessed, but no correlation was found (rho = 0.056, p = 0.53).
 |
| Figure 7. NO3-N reduction (load and %) vs NO3-N concentration (mg/L). |
 |
| Figure 8. Effect of drainage class (mw = moderately well drained, swp = somewhat poorly drained, p = poorly drained, and vp = very poorly drained) on NO3-N load reduction from CD. |
Dissolved and Total Phosphorus
Compared to nitrogen, few studies have examined the effect of DWM on DRP (n=7) and TP (n=3) concentrations and loads. Most studies reported either no change or increased P concentration following the implementation of DWM. For example, Saadat et al. (2018) found that DWM increased DRP concentration relative to free drainage in one of two field plots. Greater DRP and TP concentrations have also been observed for DWM compared to free drainage immediately after lowering the outlet to permit seasonal farming operations, but the effect on annual loads was negligible (Carstensen et al., 2019). A non-parametric Kruskal-Wallis test was used to evaluate the effect of DWM on DRP concentration and showed that DWM did not significantly impact DRP concentration compared to free drainage (p>0.05). Formal statistical testing was not completed for TP concentration given the small sample size.
Previous laboratory incubation experiments have noted that soil test P concentration, soil temperature, duration of soil saturation, and soil parent material influence the potential for P release under anaerobic soil conditions (Vadas and Sims, 1998; Plach et al., 2018). While many studies only report agronomic (0-20 cm) soil test P concentration (e.g., Mehlich-3 P, Bray-1 P), the characterization of soil P pools throughout the soil profile via sequential extractions (e.g., Hedley et al. [1982]) or using indices such as degree of phosphate saturation (DPS) or phosphate saturation index (PSI) in combination with water table measurements could serve as explanatory variables for understanding observed trends in P concentration and help identify fields where DWM may have the greatest chance of success (i.e., low potential for P concentration increase with DWM).
Phosphorus load reductions reported with controlled drainage were highly variable within and among studies. Among studies, DWM, on average, decreased DRP load by -0.11 to 0.19 kg (mean = 0.04 kg ha-1 yr-1) and TP load by 0.02 to 0.10 kg/ha/yr (mean = 0.06 kg ha-1 yr-1). Compared to free drainage, DWM, on average, decreased annual DRP load -35 to 80% (mean = 34.2%) and TP load 7 to 72% (mean = 35.5%). Kruskal-Wallis test revealed that reductions in DRP load were not significant compared to free drainage (p>0.05). It should be noted that these results should be interpreted with caution given the limited number of studies reporting P concentration and load, the large variability in P loss reductions among studies, and the sampling strategy used to quantify P load. Many of the studies included herein used weekly sampling, which often underestimates P load (Williams et al., 2015b). High-frequency sampling (e.g., daily or sub-daily) and more studies are needed to better evaluate the effectiveness of DWM on P loss.
Nearly all studies attributed the reductions in P load to decreased annual flow volumes with DWM, as changes in DRP and TP concentration were often small relative to changes in discharge. Phosphorus load reductions were often comparable to reductions in discharge (Wesstrom and Messing, 2007; Williams et al., 2015a; Tolomio and Borin, 2018), although several studies reported that P load reductions were greater than (Nash et al., 2015a) or less than (Saadat et al., 2018) reductions in discharge due to decreased or increased concentrations, respectively. Except for Sahani (2017), all studies found that, on average, controlled drainage resulted in decreased P loading in subsurface tile flow. In addition, all studies concluded that DWM may increase P loading in surface runoff; however, none of the studies included monitoring of both surface and subsurface flow components. Phosphorus concentrations are typically greater in surface runoff than subsurface tile flow (Pease et al., 2018); thus, increasing surface runoff may negate any P load reductions gained from DWM. Indeed, data from DWM field experiments in Ohio, USA, showed that surface runoff increased with the implementation of DWM, and DRP and TP loads in runoff were significantly greater at one of the two field sites evaluated (King et al., 2022). Similar to previous studies, we reiterate the need to monitor both surface and subsurface flow pathways in future DWM experiments.
Alternate Flowpaths
The question of what happens to the water that is held back in the field and does not flow through the tile drains (and the N and P transported by this water) has been raised since early studies of the practice but is rarely addressed by monitoring studies. Model simulations show that the practice likely results in some additional evapotranspiration and N uptake by plants, particularly during the growing season, as well as increased surface runoff and increased lateral and/or vertical seepage (Skaggs, 2012). The important question, as yet undetermined, is the impact of these alternate pathways on nutrient loads. The amount of denitrification is key and likely site specific. Gilliam et al. (1979) showed that in moderately well drained soil, the water table was not maintained and there was no evidence of denitrification, while in poorly drained soil, there was less reduction in discharge, but most NO3-N was denitrified.
Seepage
Seepage, either lateral or vertical, is likely the most common alternate pathway, and although denitrification could take place along the seepage pathway, this has not been well investigated. Lavaire et al. (2017) showed that lateral seepage from a field with DWM moved towards a neighboring field with free drainage, negating the perceived N reduction if only the tile discharge from the paired sites had been monitored. The magnitude of the seepage, however, has rarely been quantified. Chandrasoma et al. (2022) developed a method to quantify seepage that involved installing 33 monitoring wells and using Dupuit’s formula to estimate daily lateral seepage across each border of two plots. They showed that net seepage was into the free drainage plot and out of the DWM plot in every year. Although the amounts were relatively small compared to the amount of tile discharge, their estimated N loss reduction for DWM decreased from 80% ± 21% to 44% ± 25% when seepage was considered. Their study suggests seepage should be investigated at other sites to provide a more complete understanding of the impact of seepage on the N loss reduction of this practice.
Surface Runoff
Surface runoff is another flow pathway that may increase under DWM but is rarely monitored. Ross et al. (2016) combined measured and modeled results and found an increase of 60 mm (153%) increase in surface runoff, but an increase in NO3-N load of only 0.6 kg/ha. Drury et al. (2009) measured an increase in surface runoff from 82 mm on average in the free drainage plots to 113 mm on average in the DWM plots (2 replicates of each over 4 years), a 38% increase, but since the NO3-N concentration in the surface runoff was much lower than in the tile drains (average 2.5 mg/L), the load only increased by 0.7 kg/ha on average. The increase in surface runoff is a minor concern from the perspective of NO3-N loss, but the potential to increase erosion and P loss is more important. Tan and Zhang (2011), in a study of CD-SI, found an increase in TP loss from 0.055 kg/ha to 0.301 kg/ha, a 450% increase, but the increase from DWM is likely less. King et al. (2022) is the only study identified in this review that measured P in surface runoff with DWM from two fields with a before-after control-impact (BACI) design. They showed that the mean event surface runoff DRP and TP loads were all greater under DWM, with statistical significance at one of the two sites for each parameter. Their statistical method based on events showed statistical significance but not annual values for direct comparison with loads through other pathways, but they pointed out that their results suggest that any benefit from DWM with respect to tile DRP or TP loss could be negated in part by increased surface losses.
Tradeoffs and Limitations for Nitrogen and Phosphorus Reductions
DWM used during winter fallow periods to decrease nutrient loads results in a higher water table. As groundwater rises and intercepts shallow soils, a series of biogeochemical changes occur that can lead to greater NO3-N removal but also DRP release (fig. 9). Under saturated soil conditions, heterotrophic microbes in the soil quickly use available oxygen during normal cellular metabolism, paving the way for denitrification, a process that uses the next energetically favorable electron acceptor, NO3-N, and converts it to N2 (Carstensen et al., 2019; Gilliam, 1979). These reducing conditions also facilitate the release of P sorbed to soils rich in organic matter, iron, and with higher clay content (Trentman et al., 2020). Different forms of P are attached to and occluded within soil aggregates, which can be mobilized during periods of release that happen following snowmelt or large rain events.
 |
| Figure 9. Conceptual drawing of DWM in winter conditions, including expected changes in organic matter, temperature, pore water NO3-N, soil phosphorus, and redox. Shading indicates areas of greater uncertainty. |
Within the soil profile, properties change with depth, such that there is higher and potentially more variable organic matter in soils near the surface, and that amount generally decreases with depth. (fig. 9). This same pattern can be observed with soil P and concentrations of NO3-N in pore water. In temperate climates during winter, these surface soils are also often frozen, with temperatures increasing with depth. This complex interaction of soil organic matter, nutrients, redox conditions, and temperature can lead to greater NO3-N removal and P release in agricultural fields with DWM, and this shifts as water tables are managed at the control structure near the outlet. While there is considerable knowledge of these mechanisms in wetlands and floodplains (Noe et al., 2013; Jones et al., 2014), to our knowledge, no studies have examined these environmental drivers in a systematic way to test these hypotheses. Incorporation of process-based measurements (e.g., microbial metabolism, denitrification) and discretization of the different P pools present in the soil should be incorporated into future studies to better understand these tradeoffs and place practices in locations that have the highest potential for net benefit.
Our synthesis highlights a wide range of NO3-N and P load reductions and that these are largely driven by changes in discharge volume, which is a combination of the impact of DWM but also climate, field management practices (e.g., fertilizer application, soil test P concentration), soil characteristics (e.g., topography, soil parent material), and differences in the outlet elevation within the control structure. These interactions make paired comparisons challenging; however, this work and others underscore the importance of measuring and collecting information on environmental factors that may control performance so that the variability among and within sites can be better constrained.
Cost-Effectiveness
Crop Yield Effects
The potential for DWM to increase crop yields is an important factor in its cost-effectiveness. At least 15 studies have published the impact of drainage water management on corn and/or soybean yields (table 2). Although many found no significant effect, seven studies have shown increases up to 18% and two showed a decrease, one of which was for only one year.
Yield effects are expected to increase as automated control structures, which provide the potential for adaptive management to maintain optimal water levels, become more widely used (Youssef et al., 2023). These systems include programmable valves triggered by water level sensors to hold or release water (Miller et al., 2022), and add about $5000 to the cost of each control structure. However, no published studies have examined the effects of DWM automated structures on crop yield, so our economic analysis considers a yield increase of 0% to 3% using manually-adjusted structures only.
Costs
Previous Analyses of Drainage Water Management Costs
Nistor and Lowenberg-Deboer (2007) developed a farm level simulation analysis to examine the yield changes that would be required to break-even when installing a DWM system. Their approach assumed that the existing drainage system is in place such that the only change to the system is the installation of the control structures. They considered a 1215 ha corn-soybean farm in Indiana, with slopes such that each structure could control 8 ha, requiring 150 total structures. Their model includes constraints to control the time available to manage and operate the control structures. This approach is useful because it addresses the tradeoffs farmers will face when they have a DWM system, such as taking the time to manage the drainage water system rather than some other part of the farm operation at a critical time. They found that installation of the DWM system achieves a break-even point when yields are 2% greater as a result of the DWM system with a subsidy from a conservation program, and 4% without the subsidy. However, their results also imply that under some circumstances (i.e., weather conditions), it would take more than a 10% yield gain to achieve break-even. They speculate that these types of conditions, while only occurring once in a while, are an important impediment for farmers who are considering installation. Estimates from Nistor and Lowenberg-Deboer indicate that costs are $125 ha-1 to install if there is a 50% government subsidy and $255 ha-1 to install without the subsidy. Annual opportunity costs of management are $9 ha-1 when converted to a present value.
Crabbe et al. (2012) examined the costs of implementing DWM systems in Southern Ontario, Canada. Their cost estimates are given in 2006 Canadian $, so we translated these to 2022 US $ by first converting them to 2006 US $ using an exchange rate of 1.15C$ per 1US$, and then adjusting for inflation using the all-commodity Producer Price Index, which means we multiply by 1.50 in this case. They calculate that the costs are $39/ha/yr to install and manage the system. Using their interest rate of 6%, and assuming a 20-year life, this is $360/structure for installation and management. In the Crabbe et al. (2012) model, each control structure controls only 4 ha. Based on their paired field comparison, the yield increase was 3.3%-3.8% for corn and soybeans. Given prices observed at the time, the revenue increase due to the DWM system was found to be $50/ha/yr for soybean and $80/ha/yr for corn, suggesting that costs outweigh benefits by around 2 to 1.
| Table 2. Published impacts of drainage water management on corn and/or soybean yields. Statistically significant results as noted in the papers are shown in bold and noted with *. | |||
| Study | Location | Corn Yield Effect | Soybean Yield Effect |
| Tan et al. (1998) | Woodslee, Ontario | No effect | No effect |
| Fausey (2005) | Wood Cty, OH | No effect | No effect |
| Wesstrom and Messing (2007) | Sweden | 2%-18% | |
| Drury et al. (2009) | Woodslee, Ontario | No effect | No effect |
| Smith and Kellman (2011) | Nova Scotia | -14%, 1 year only | |
| Jaynes (2012) | Story Cty, IA | No effect | 8%* |
| Delbecq et al. (2012) | Randolph Cty, IN | 3.3%* | |
| Crabbe et al. (2012) | Ontario, Canada | 3.3%* | 3.8%* |
| Helmers et al. (2012) | Washington Cty, IA | Decrease* | No effect |
| Ghane et al. (2012) | 7 sites in Ohio | 3.3%* | 2.1%* |
| Poole et al. (2013) | Plymouth, NC | 11%* | 10%* |
| Schott et al. (2017) | Washington Cty, IA | No effect | No effect |
| Smith et al. (2019) | Nova Scotia | 7% | |
| Youssef et al. (2023) | 13 sites in 6 Midwestern states and NC | No effect | No effect |
| Baird et al. (2024) | Randolph Cty, IN | 2.3%* | No effect |
Christianson et al. (2013) reviewed several different practices and presented methods to calculate the costs of implementing them to reduce N losses. They calculated the installation costs for DWM at $147 ha-1 to $591 ha-1, with the lowest costs occurring if the system is installed as part of a new drainage system at 1 structure per 8 ha and the higher costs at 1 structure per 4 ha. Annual maintenance costs focus on changing the gates every 8 years, and annual costs vary from $18 ha-1 to $35 ha-1. They also included costs to manage the system annually of $1-5 ha-1., for total costs of $235 ha-1 to $925 ha-1 (table 3).
| Table 3. Range of installation and management costs across studies reviewed (USD 2022). | ||
| Present value of total cost ($ ha-1) | Annual value ($ ha-1 yr-1) | |
| Nistor and Lowenberg-Deboer (2007) | $329 | $38 |
| Crabbe et al. (2012) | $449 | $39 |
| Christianson et al. (2013) | $235-$925 | $12-$47 |
Model of Costs
The costs for installation and management of DWM vary substantially in the literature due to a range of factors, including the field characteristics and specific design used by individual researchers, the installation practices in the region where the study was conducted, and other site or individual specific characteristics. We have conducted a sensitivity analysis across various input assumptions using recent estimates of hardware costs obtained from drainage companies in the Midwestern U.S.
The typical installation is assumed to occur in a field in which one structure can control 8 ha with a structure depth of 1.2 m and a lifespan of 20 years (lifespan for “structure for water control” in USDA-NRCS, 2022a). Although the area controlled by one structure varies widely, 4 to 8 ha were used in the USDA-NRCS Practice Scenarios for Fiscal Year 2023 for Indiana (USDA-NRCS, 2022b) and other states. We assume the installation occurs in a field that can be retrofit with the control structures. Under typical assumptions used in this example, installation costs are estimated at $220 ha-1, and the present value of management and maintenance costs are estimated at $69 ha-1, for a total cost of $289 ha-1, or $23/ha/yr when spread over 20 years assuming an interest rate of 5%. Under the high cost assumptions, which entail fewer hectares controlled per structure (4 ha, perhaps due to a greater field slope) and a larger control structure (depth and pipe size), costs could be as much as $80 ha-1 yr-1. Under the low-cost assumptions, which would occur in fields with low slope and more hectares controlled per structure, costs could be as low as $12 ha-1 yr-1.
This enormous range of costs illustrates how important site-specific factors will influence the potential cost of installation and operation on any given farm. The factor that causes the greatest variation in costs is the number of acres that can be controlled by each structure. Sites with greater slopes are likely to allow for less land to be controlled per structure, which has a large effect on cost. Labor needs and miscellaneous costs during installation also have a relatively important effect on costs. The other factors are important but affect costs by only 10%-15%. Differences in assumptions about the annual time to manage the structures have a relatively modest effect.
Under the typical installation, and a corn price of $236 t-1 ($6 bu-1) and soybean price of $551 t-1 ($15 bu-1), it would take a 1.0% increase in yield as a result of the DWM project to breakeven. Alternatively, cost share programs could offset some of the costs of installing and operating a system. Assuming a 3.0% increase in yields, consistent with existing studies, farmers could achieve a net benefit of $50/ha/yr by installing this practice. Many farmers are likely to be concerned that yield increases may not be achieved in all circumstances, and thus need a subsidy to encourage the installation of this practice. In this case, a cost share payment of $1784 per structure, or $18 ha-1 yr-1, would cover the installation cost. An additional payment of $6 ha-1 yr-1 would cover annual maintenance and management costs.
Cost Effectiveness ($/kg Removal)
Cost effectiveness, as $/kg N or P removed, is calculated by dividing the cost in $/ha by the nutrient loss reduction in kg/ha. This calculation assumes that cost is not directly linked to nutrient loss reduction, but rather to field-specific characteristics, such as topography, which determine the number of structures required. Given the range of both costs and nutrient reductions in the literature, the range of cost effectiveness is high. For a range of cost, we used the lower and upper numbers in table 4, while for N loss reduction, we used the confidence interval from the mixed effects model for the mean described above. The resulting cost-effectiveness for yield increases of 0% and 3% is provided in table 5. With no yield benefits, the cost of nitrogen ranges from $1 to $11 kg-1 N removed, with a typical cost of $2. With a 3% yield increase, the practice results in a net increase in income per hectare, meaning that there is a negative “cost-effectiveness” (net income gain) for the low and typical cost scenario, although there is still a cost of $1kg-1 N removed for the high cost scenario.
| Table 4. Installation and management/maintenance costs for a “typical” Midwestern setup, a high cost set up and a low cost setup, with a 20-year lifetime.[a] | ||||
| Inputs and Costs | Units | Low Cost | Typical | High Cost |
| Area controlled by each structure (Flat topography (i.e., < 1% slope) allows higher value.) | ha/structure | 10 (25 acres) | 8 (20 acres) | 4 (10 acres) |
| Structure depth | meters | 1.2 | 1.2 | 1.2 |
| Pipe size (drainage pipe on which structure is mounted) | cm | 15.2 | 20.3 | 25.4 |
| Installation labor needs | Hours/structure | 3 | 4 | 5 |
| Extra piping | Meters/structure | 3.03 | 7.6 | 15.2 |
| Miscellaneous costs | $/structure | 0 | 200 | 400 |
| Labor/Opportunity Costs | $ hr-1 | $12 | $18 | $25 |
| Time to manage | hrs/adjustment/structure | 0.50 | 0.50 | 0.75 |
| Number of adjustments/year | times yr-1 | 2 | 4 | 8 |
| Maintenance costs | % initial capital yr-1 | 0.25% | 0.5% | 0.75 |
| Installation | $ ha-1 | $128 | $220 | $631 |
| Management and maintenance (present value) | ($ ha-1) | $19 | $69 | $367 |
| Total costs (present value) | ($ ha-1) | $147 | $289 | $998 |
| Total costs | ($ ha-1 yr-1) | $12 ha-1 yr- | $24 ha-1 yr- | $80 ha-1 yr- |
| Adjusted Costs/Benefits with a potential 3% yield increase (base yield 14.1 t ha-1 for corn and 3 t ha-1 for soybean) | ||||
| Gross costs due to yield gain (negative means an increase in income) | $ ha-1yr-1 | ($73) | ($73) | ($73) |
| Net costs including effect of yield increase (negative means an increase in income)” | $ ha-1yr-1 | ($61) | ($59) | $7 |
[a] Interest rate = 5%; Corn price =$236/t; Soybean price = $551/t | ||||
| Table 5. Low, median, and high values for cost-effectiveness, based on cost and nitrogen reductions. Negative values represent a net gain in income. | ||||
| Low | Typical | High | ||
| Cost, including potential yield benefits ($/ha/yr) | 0% yield increase | $12 | $24 | $80 |
| 3% yield increase | ($61) | ($50) | $7 | |
| Nitrogen reduction (kg/ha/yr), based on mean +/- confidence interval | 7.4 | 13.3 | 17.4 | |
| Cost effectiveness ($/kg N removed) | 0% yield increase | $1[a] | $2 | $11[b] |
| 3% yield increase | ($3)[a] | ($3) | $1[b] | |
[a] Calculated by dividing the low cost by the high nitrogen reduction and rounded to whole dollar to reflect uncertainty. [b] Calculated by dividing the high cost by the low nitrogen reduction and rounded to whole dollar to reflect uncertainty. | ||||
Monitoring Recommendations
Monitoring Considerations
Drain discharge at a daily or higher frequency is key to understanding DWM performance. Monitoring drain discharge in subsurface tile drainage systems is challenging because standard flow measurements, including weirs or flumes, do not work in pipes in which the outlet is sometimes submerged, and may even have backward flow depending on outlet conditions. All methods are vulnerable to instrumentation breakdowns, and therefore we recommend implementing duplicate methods to monitor flow because of its importance to understanding practice effectiveness. Monitoring the water table is also key for understanding the effects of the practice on water table depth.
Gaining an understanding of the complete water budget is a critical need in future studies. Surface runoff and seepage (lateral and vertical) have been rarely monitored, but are needed to determine DWM effectiveness as a nutrient reduction practice. Monitoring surface runoff is challenging because it may require changing field topography for collection and because of the highly episodic nature of runoff. Chandrasoma et al. (2022) demonstrated an effective method for monitoring lateral seepage that can guide future researchers, although they point out that their method of estimating seepage at the field boundary did not consider seepage directly around the control structure, or vertical seepage, which can also be important pathways bypassing tile drains.
Nutrient sampling and analysis must be carried out at a frequency adequate for calculating loads. While grab samples have sometimes been assumed to be enough for NO3-N monitoring, Rozemeijer et al. (2016) showed the high and frequent variability and the key role it plays even for NO3-N monitoring. Phosphorus monitoring requires even more frequent samples to capture accurate loads, as concentration peaks sharply near the beginning of a rainfall event and most P is transported during a few large storm events each year (King et al., 2015). Most (6 of 8 studies in the P review) used either a weekly grab or a weekly composite sampling strategy for collecting water samples. Increasing the frequency of nutrient analysis and measuring mass of P stored in different forms in the soil may provide additional insights on DWM effects and tradeoffs (Williams et al., 2015b). Because of the complex interactions in management, soil properties, climate, and others, future studies should include several things: (1) measurements of microbial activity, including denitrification and metabolism, (2) environmental parameters such as different forms of soil P, organic matter, infiltration rate, and redox potential, and (3) clear and consistent documentation of land management practices and DWM design. Inclusion of these in future studies will aid decision makers in design and management considerations (e.g., tile spacing, outlet depth management, timing of retaining water) and support the implementation of this practice in locations where it will have the greatest water quality benefit.
Analysis Considerations
The statistical method used for determining the effectiveness of DWM can also contribute to variability in load reductions. The two most common approaches used in the reviewed studies were either calculating the difference in annual load between free drainage and DWM fields, or a paired-field approach using a before-after control-impact (BACI) study design. These methods may result in different load reductions. For example, Carstensen et al. (2019) used a BACI approach and found that controlled drainage decreased TP load by -57 to 46% (average = 7%). However, comparing differences in annual TP load between their fields with free drainage and DWM shows that DWM decreased TP load by -50 to 79% (average = 19%; see table 2). While BACI is considered a more rigorous approach, adding a year or more of data collection adds an additional barrier to monitoring the practice. Shedekar et al. (2021) examined the differences due to the analysis method in depth, using data published by Williams et al. (2015a) and recommended multiple analysis approaches to better understand the practice. Analyses during different management periods may also add insight; for example, Saadat et al. (2018) separated the lower level of control during the growing season from the higher level of control during the non-growing season.
Documentation Considerations
In addition to drain flow, water table, nutrient concentrations, and loads that are commonly reported, DWM studies should report the outlet control height throughout the year, as well as the drainage design parameters and soil characteristics needed to calculate the three key parameters suggested by Skaggs (2017): drainage intensity, drainage coefficient, and the Kirkham coefficient. Providing data in a publicly-available dataset will increase the re-analysis possible for these expensive and intensive monitoring studies. The Transforming Drainage dataset (Chigladze et al., 2021) provides an example.
Conclusion
Nutrient load reduction from drainage water management varied widely, both by site and by individual years at each site. The overall mean nitrogen load reductions of the 31 sites that met our selection criteria was 13.3 kg/ha with a confidence interval of 9.4 to 17.3 kg/ha, or a 46% load reduction efficiency. Nutrient reduction is due mainly to a reduction in flow, as few reductions in concentration have been found or reported. Drainage water management decreased annual DRP and TP load compared to free drainage by an average of 0.04 kg/ha (34%) and 0.06 kg/ha (36%), respectively. Findings indicate that DWM can be an effective practice for decreasing nutrient loading from drained landscapes; however, trade-offs may increase nutrient loss through seepage and surface runoff.
Economic analysis showed that the cost of N reduction would range from $1 to $11 per kg N removal if there was no yield benefit, while with a 3% yield increase, there is likely to be a net economic benefit to the producer. Monitoring of the complete water budget, together with geochemical parameters that would allow quantification of denitrification, is needed to improve our understanding of the overall benefits of the practice.
Supplemental Material
The supplemental information mentioned in this article is available for download from the ASABE Figshare repository at: https://doi.org/10.13031/23596572
Acknowledgments
This material is based upon work that was supported by the National Institute of Food and Agriculture, the U.S. Department of Agriculture (award number 2015-68007-23193), and the U.S. Environmental Protection Agency (award number X7-00E02782) through the Indiana State Department of Agriculture.
References
Abendroth, L. J., Chighladze, G., Frankenberger, J. R., Bowling, L. C., Helmers, M. J., Herzmann, D. E.,... Youssef, M. (2022). Paired field and water measurements from drainage management practices in row-crop agriculture. Sci. Data, 9(1), 257. https://doi.org/10.1038/s41597-022-01358-7
Adeuya, R., Utt, N., Frankenberger, J., Bowling, L., Kladivko, E., Brouder, S., & Carter, B. (2012). Impacts of drainage water management on subsurface drain flow, nitrate concentration, and nitrate loads in Indiana. J. Soil Water Conserv., 67(6), 474-484. https://doi.org/10.2489/jswc.67.6.474
Agricultural Drainage Management Coalition (ADMC). (2012). Drainage water management for midwestern agriculture. A final report for USDA-NRCS Conservation Innovation Grant 68-3A75-6-116.
Ayars, J. E., Christen, E. W., & Hornbuckle, J. W. (2006). Controlled drainage for improved water management in arid regions irrigated agriculture. Agric. Water Manag., 86(1), 128-139. https://doi.org/10.1016/j.agwat.2006.07.004
Baird, A., Frankenberger, J., Bowling, B., & Kladivko, E. (2024). Impact of controlled drainage on crop yield including within-field variability. J. ASABE67(3), 717-727, https://doi.org/10.13031/ja.15520
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Softw., 67(1), 1-48. https://doi.org/10.18637/jss.v067.i01
Carstensen, M. V., Børgesen, C. D., Ovesen, N. B., Poulsen, J. R., Hvid, S. K., & Kronvang, B. (2019). Controlled drainage as a targeted mitigation measure for nitrogen and phosphorus. J. Environ. Qual., 48(3), 677-685. https://doi.org/10.2134/jeq2018.11.0393
Carstensen, M. V., Hashemi, F., Hoffmann, C. C., Zak, D., Audet, J., & Kronvang, B. (2020). Efficiency of mitigation measures targeting nutrient losses from agricultural drainage systems: A review. Ambio, 49(11), 1820-1837. https://doi.org/10.1007/s13280-020-01345-5
Chandrasoma, J., Christianson, R. D., Davidson, P. C., Cooke, R. A., & Christianson, L. E. (2022). Seepage-corrected N loss reduction assessment for the practice of controlled drainage in Illinois. Proc. 11th Int. Drainage Symp. (pp. 1-7). St. Joseph, MI: ASABE. https://doi.org/10.13031/ids.202200019
Chighladze, G., Abendroth, L. J., Herzmann, D., Helmers, M. J., Ahiablame, L., Allred, B.,... Youssef, M. (2021). Transforming drainage research data (USDA-NIFA Award No 2015-68007-23193). Ag Data Commons. https://doi.org/10.15482/USDA.ADC/1521092
Christianson, L., Tyndall, J., & Helmers, M. (2013). Financial comparison of seven nitrate reduction strategies for Midwestern agricultural drainage. Water Resour. Econ., 2-3, 30-56. https://doi.org/10.1016/j.wre.2013.09.001
Clayton, B. C., & Jones, L. A. (1941). Controlled drainage in the northern Everglades of Florida. Agric. Eng.
Cooke, R., & Verma, S. (2012). Performance of drainage water management systems in Illinois, United States. J. Soil Water Conserv., 67(6), 453-464. https://doi.org/10.2489/jswc.67.6.453
Crabbé, P., Lapen, D. R., Clark, H., Sunohara, M., & Liu, Y. (2012). Economic benefits of controlled tile drainage: Watershed evaluation of beneficial management practices, South Nation River basin, Ontario. Water Qual. Res. J., 47(1), 30-41. https://doi.org/10.2166/wqrjc.2012.007
Delbecq, B. A., Brown, J. P., Florax, R. J., Kladivko, E. J., Nistor, A. P., & Lowenberg-DeBoer, J. M. (2012). The impact of drainage water management technology on corn yields. Agron. J., 104(4), 1100-1109. https://doi.org/10.2134/agronj2012.0003
Doty, C. W. (1980). Crop water supplied by controlled and reversible drainage. Trans. ASAE, 23(5), 1122-1126. https://doi.org/10.13031/2013.34731
Drury, C. F., Tan, C. S., Reynolds, W. D., Welacky, T. W., Oloya, T. O., & Gaynor, J. D. (2009). Managing tile drainage, subirrigation, and nitrogen fertilization to enhance crop yields and reduce nitrate loss. J. Environ. Qual., 38(3), 1193-1204. https://doi.org/10.2134/jeq2008.0036
Evans, R. O., Gilliam, J. W., & Skaggs, R. W. (1989). Effects of agricultural water table management on drainage water quality. WRRI Project No. 70056/70081. Retrieved from https://repository.lib.ncsu.edu/handle/1840.4/1856
Fausey, N. R. (2005). Drainage management for humid regions. Int. Agric. Eng. J., 14(4), 209-214.
Feyereisen, G. W., Hay, C. H., Christianson, R. D., & Helmers, M. J. (2022). Frontier: eating the metaphorical elephant: meeting nitrogen reduction goals in upper Mississippi River Basin states. J. ASABE65(3), 621–631, https://doi.org/10.13031/ja.14887
Fouss, J. L., & Sullivan, M. (2012). Agricultural Drainage Management Systems Task Force (ADMSTF). In World environmental and water resources congress 2009: Great rivers (pp. 1-10). https://doi.org/10.1061/41036(342)411
Frankenberger, J., Kladivko, E., Bowling, L., Helmers, M., Kjaersgaards, J., McMaine, J.,... Youssef., M. (2023). Questions and answers about controlled drainage for the Midwest. In press. Purdue University Extension publication ABE-166.
Ghane, E., Fausey, N. R., Shedekar, V. S., Piepho, H. P., Shang, Y., & Brown, L. C. (2012). Crop yield evaluation under controlled drainage in Ohio, United States. J. Soil Water Conserv., 67(6), 465-473. https://doi.org/10.2489/jswc.67.6.465
Gilliam, J. W., Skaggs, R. W., & Weed, S. B. (1979). Drainage control to diminish nitrate loss from agricultural fields. J. Environ. Qual., 8(1), 137-142. https://doi.org/10.2134/jeq1979.00472425000800010030x
Hedley, M. J., Stewart, J. W., & Chauhan, B. S. (1982). Changes in inorganic and organic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Sci. Soc. Am. J., 46(5), 970-976. https://doi.org/10.2136/sssaj1982.03615995004600050017x
Helmers, M. J., Abendroth, L., Reinhart, B., Chighladze, G., Pease, L., Bowling, L., Youssef, M., Ghane, E., Ahiablame, L., Brown, L. and Fausey, N. (2022). Impact of controlled drainage on subsurface drain flow and nitrate load: A synthesis of studies across the U.S. Midwest and Southeast. Agric. Water Manag., 259, 107265. https://doi.org/10.1016/j.agwat.2021.107265
Helmers, M., Christianson, R., Brenneman, G., Lockett, D., & Pederson, C. (2012). Water table, drainage, and yield response to drainage water management in southeast Iowa. J. Soil Water Conserv., 67(6), 495-501. https://doi.org/10.2489/jswc.67.6.495
Jaynes, D. B. (2012). Changes in yield and nitrate losses from using drainage water management in central Iowa, United States. J. Soil Water Conserv., 67(6), 485-494. https://doi.org/10.2489/jswc.67.6.485
Jones, C. N., Scott, D. T., Edwards, B. L., & Keim, R. F. (2014). Perirheic mixing and biogeochemical processing in flow-through and backwater floodplain wetlands. Water Resour. Res., 50(9), 7394-7405. https://doi.org/10.1002/2014WR015647
Jongedyk, H. A., Hickok, R. B., & Mayer, I. D. (1954). Changes in drainage properties of a muck soil as a result of drainage practices. Soil Sci. Soc. Am. J., 18(1), 72-76. https://doi.org/10.2136/sssaj1954.03615995001800010018x
Kesicka, B., Stasik, R., & Kozlowski, M. (2022). Effects of modelling studies on controlled drainage in agricultural land on reduction of outflow and nitrate losses – a meta-analysis. PLoS One, 17(4), e0267736. https://doi.org/10.1371/journal.pone.0267736
King, K. W., Hanrahan, B. R., Stinner, J., & Shedekar, V. S. (2022). Field scale discharge and water quality response, to drainage water management. Agric. Water Manag., 264, 107421. https://doi.org/10.1016/j.agwat.2021.107421
King, K. W., Williams, M. R., & Fausey, N. R. (2015). Contributions of systematic tile drainage to watershed-scale phosphorus transport. J. Environ. Qual., 44(2), 486-494. https://doi.org/10.2134/jeq2014.04.0149
Lavaire, T., Gentry, L. E., David, M. B., & Cooke, R. A. (2017). Fate of water and nitrate using drainage water management on tile systems in east-central Illinois. Agric. Water Manag., 191, 218-228. https://doi.org/10.1016/j.agwat.2017.06.004
Miller, S. A., Witter, J. D., & Lyon, S. W. (2022). The impact of automated drainage water management on groundwater, soil moisture, and tile outlet discharge following storm events. Agric. Water Manag., 272, 107828. https://doi.org/10.1016/j.agwat.2022.107828
Nash, P. R., Nelson, K. A., Motavalli, P. P., Nathan, M., & Dudenhoeffer, C. (2015a). Reducing phosphorus loss in tile water with managed drainage in a claypan soil. J. Environ. Qual., 44(2), 585-593. https://doi.org/10.2134/jeq2014.04.0146
Nash, P., Nelson, K., & Motavalli, P. (2015b). Reducing nitrogen loss with managed drainage and polymer-coated urea. J. Environ. Qual., 44(1), 256-264. https://doi.org/10.2134/jeq2014.05.0238
Nistor, A. P., & Lowenberg-DeBoer, J. (2007). Drainage water management impact on farm proftability. J. Soil Water Conserv., 62(6), 443-446. Retrieved from https://www.jswconline.org/content/jswc/62/6/443.full.pdf
Noe, G. B., Hupp, C. R., & Rybicki, N. B. (2013). Hydrogeomorphology influences soil nitrogen and phosphorus mineralization in floodplain wetlands. Ecosystems, 16(1), 75-94. https://doi.org/10.1007/s10021-012-9597-0
Pease, L. A., King, K. W., Williams, M. R., LaBarge, G. A., Duncan, E. W., & Fausey, N. R. (2018). Phosphorus export from artificially drained fields across the Eastern Corn Belt. J. Great Lakes Res., 44(1), 43-53. https://doi.org/10.1016/j.jglr.2017.11.009
Plach, J. M., Macrae, M. L., Williams, M. R., Lee, B. D., & King, K. W. (2018). Dominant glacial landforms of the lower Great Lakes region exhibit different soil phosphorus chemistry and potential risk for phosphorus loss. J. Great Lakes Res., 44(5), 1057-1067. https://doi.org/10.1016/j.jglr.2018.07.005
Poole, C.A., Skaggs, R.W., Cheschier, G.M., Youssef, M.A. & Crozier, C.R. (2013). Effects of drainage water management on crop yields in North Carolina. Journal of soil and water conservation, 68(6), 429-437.
Poole, C. A., Skaggs, R. W., Youssef, M. A., Chescheir, G. M., & Crozier, C. R. (2018). Effect of drainage water management on nitrate nitrogen loss to tile drains in North Carolina. Trans. ASABE, 61(1), 233-244. https://doi.org/10.13031/trans.12296
Ramoska, E., Bastiene, N., & Saulys, V. (2011). Evaluation of controlled drainage efficiency in Lithuania. Irrig. Drain., 60(2), 196-206. https://doi.org/10.1002/ird.548
Ross, J. A., Herbert, M. E., Sowa, S. P., Frankenberger, J. R., King, K. W., Christopher, S. F.,... Yen, H. (2016). A synthesis and comparative evaluation of factors influencing the effectiveness of drainage water management. Agric. Water Manag., 178, 366-376. https://doi.org/10.1016/j.agwat.2016.10.011
Rozemeijer, J. C., Visser, A., Borren, W., Winegram, M., Van Der Velde, Y., Klein, J. & Broers, H. P., 2016. High-frequency monitoring of water fluxes and nutrient loads to assess the effects of controlled drainage on water storage and nutrient transport. Hydrology and Earth System Sciences, 20(1), 347-358.
Saadat, S., Bowling, L., Frankenberger, J., & Kladivko, E. (2018). Nitrate and phosphorus transport through subsurface drains under free and controlled drainage. Water Res., 142, 196-207. https://doi.org/10.1016/j.watres.2018.05.040
Sahani, A. (2017). A demonstration study of drainage water management in Eastern South Dakota. MS thesis. South Dakota State University. Retrieved from https://openprairie.sdstate.edu/etd/2148
Satchithanantham, S., Ranjan, R. S., & Bullock, P. (2014). Protecting water quality using controlled drainage as an agricultural BMP for potato production. Trans. ASABE, 57(3), 815-826. https://doi.org/10.13031/trans.57.10385
Schott, L., Lagzdins, A., Daigh, A. L., Craft, K., Pederson, C., Brenneman, G., & Helmers, M. J. (2017). Drainage water management effects over five years on water tables, drainage, and yields in southeast Iowa. J. Soil Water Conserv., 72(3), 251-259. https://doi.org/10.2489/jswc.72.3.251
Sharma, A. (2018). Application of drainage water management and saturated buffers for conservation drainage in South Dakota. MS thesis. Agricultural and Biosystems Engineering, South Dakota State University.
Shedekar, V. S., King, K. W., Fausey, N. R., Islam, K. R., Soboyejo, A. B., Kalcic, M. M., & Brown, L. C. (2021). Exploring the effectiveness of drainage water management on water budgets and nitrate loss using three evaluation approaches. Agric. Water Manag., 243, 106501. https://doi.org/10.1016/j.agwat.2020.106501
Skaggs, R. W. (2017). Coefficients for quantifying subsurface drainage rates. Appl. Eng. Agric., 33(6), 793-799. https://doi.org/10.13031/aea.12302
Skaggs, R. W., Fausey, N. R., & Evans, R. O. (2012). Drainage water management. J. Soil Water Conserv., 67(6), 167A-172A. https://doi.org/10.2489/jswc.67.6.167A
Smith, D. R., White, M., McLellan. E., L., Pampell, R., & Harmel, R. D. (2019). Using the Conservation Practice Effectiveness (CoPE) database to assess adoption tradeoffs. J. Soil Water Conserv., 74(6), 554-559. https://doi.org/10.2489/jswc.74.6.554
Smith, E. L., & Kellman, L. M. (2011). Nitrate loading and isotopic signatures in subsurface agricultural drainage systems. J. Environ. Qual., 40(4), 1257-1265. https://doi.org/10.2134/jeq2010.0489
Stephens, J. C. (1955). Drainage of peat and muck lands. In Water. The 1955 yearbook of agriculture. Washington, DC: USDA-ARS. Retrieved from https://naldc.nal.usda.gov/download/IND43894609/PDF.
Strock, J., Reinhart, B., & Frankenberger, J. (2022). Transforming drainage site summary for Minnesota-Redwood. Retrieved from https://transformingdrainage.org/wp-content/uploads/2022/05/MN_Redwood1_layout.pdf
Sunohara, M. D., Craiovan, E., Topp, E., Gottschall, N., Drury, C. F., & Lapen, D. R. (2014). Comprehensive nitrogen budgets for controlled tile drainage fields in Eastern Ontario, Canada. J. Environ. Qual., 43(2), 617-630. https://doi.org/10.2134/jeq2013.04.0117
Tan, C. S., & Zhang, T. Q. (2011). Surface runoff and sub-surface drainage phosphorus losses under regular free drainage and controlled drainage with sub-irrigation systems in southern Ontario. Can. J. Soil Sci., 91(3), 349-359. https://doi.org/10.4141/cjss09086
Tan, C. S., Drury, C. F., Soultani, M., van Wesenbeeck, I. J., Ng, H. Y., Gaynor, J. D., & Welacky, T. W. (1998). Effect of controlled drainage and tillage on soil structure and tile drainage nitrate loss at the field scale. Water Sci. Technol., 38(4-5), 103-110. https://doi.org/10.2166/wst.1998.0593
Tolomio, M., & Borin, M. (2018). Water table management to save water and reduce nutrient losses from agricultural fields: 6 years of experience in North-Eastern Italy. Agric. Water Manag., 201, 1-10. https://doi.org/10.1016/j.agwat.2018.01.009
Trentman, M. T., Tank, J. L., Jones, S. E., McMillan, S. K., & Royer, T. V. (2020). Seasonal evaluation of biotic and abiotic factors suggests phosphorus retention in constructed floodplains in three agricultural streams. Sci. Total Environ., 729, 138744. https://doi.org/10.1016/j.scitotenv.2020.138744
USDA-NRCS. (2022a). National Handbook of Conservation Practices, Chapter 3. USDA-NRCS. Retrieved from https://directives.sc.egov.usda.gov/OpenNonWebContent.aspx?content=48740.wba
USDA-NRCS. (2022b). Indiana Practice Scenarios - Fiscal Year 2023. USDA-NRCS. Retrieved from https://www.nrcs.usda.gov/sites/default/files/2022-11/Indiana-Scenarios-23-payment-rates.pdf
USDA-NRCS Conservation Practice Standards. (2020). 554: Drainage water management. Washington, DC: USDA-NRCS. Retrieved from https://www.nrcs.usda.gov/sites/default/files/2022-09/Drainage_Water_Management_554_CPS_10_2020.pdf
Vadas, P. A., & Sims, J. T. (1998). Redox status, poultry litter, and phosphorus solubility in Atlantic Coastal Plain soils. Soil Sci. Soc. Am. J., 62(4), 1025-1034. https://doi.org/10.2136/sssaj1998.03615995006200040025x
Wang, Z., Shao, G., Lu, J., Zhang, K., Gao, Y., & Ding, J. (2020). Effects of controlled drainage on crop yield, drainage water quantity and quality: A meta-analysis. Agric. Water Manag., 239, 106253. https://doi.org/10.1016/j.agwat.2020.106253
Wesström, I., & Messing, I. (2007). Effects of controlled drainage on N and P losses and N dynamics in a loamy sand with spring crops. Agric. Water Manag., 87(3), 229-240. https://doi.org/10.1016/j.agwat.2006.07.005
Willardson, L. S., Meek, B. D., Grass, L. B., Dickey, L. G., & Bailey, J. W. (1972). Nitrate reduction with submerged drains. Trans. ASAE, 15(1), 84-85. https://doi.org/10.13031/2013.37836
Williams, M. R., King, K. W., & Fausey, N. R. (2015a). Drainage water management effects on tile discharge and water quality. Agric. Water Manag., 148, 43-51. https://doi.org/10.1016/j.agwat.2014.09.017
Williams, M. R., King, K. W., Macrae, M. L., Ford, W., Van Esbroeck, C., Brunke, R. I.,... Schiff, S. L. (2015b). Uncertainty in nutrient loads from tile-drained landscapes: Effect of sampling frequency, calculation algorithm, and compositing strategy. J. Hydrol., 530, 306-316. https://doi.org/10.1016/j.jhydrol.2015.09.060
Woli, K. P., David, M. B., Cooke, R. A., McIsaac, G. F., & Mitchell, C. A. (2010). Nitrogen balance in and export from agricultural fields associated with controlled drainage systems and denitrifying bioreactors. Ecol. Eng., 36(11), 1558-1566. https://doi.org/10.1016/j.ecoleng.2010.04.024
Youssef, M. A., Strock, J., Bagheri, E., Reinhart, B. D., Abendroth, L. J., Chighladze, G., Ghane, E., Shedekar, V., Fausey, N. R., Frankenberger, J. R. & Helmers, M. J. (2023). Impact of controlled drainage on corn yield under varying precipitation patterns: A synthesis of studies across the U.S. Midwest and Southeast. Agric. Water Manag., 275, 107993. https://doi.org/10.1016/j.agwat.2022.107993