A review of ecosystem services from edge-of-field practices in tile-drained agricultural systems in the United States Corn Belt Region

Abstract

Edge-of-field management practices that reduce nutrient pollution from tile drainage while contributing habitat and other ecosystem services are needed to enhance agricultural systems in the US Corn Belt Region. In this review, we identified edge-of-field and catchment scale agricultural conservation practices for intercepting and treating tile drainage. The reviewed conservation practices were (1) controlled drainage, also known as drainage water management (USDA-NRCS Code 554); (2) drainage water recycling (USDA-NRCS Code 447); (3) denitrifying bioreactors (USDA-NRCS Code 605); (4) saturated buffers (USDA-NRCS Code 604); and (5) constructed or restored wetlands designed for water quality improvement (USDA-NRCS Code 656) herein referred to as water quality wetlands. We examined 119 studies that had information on one or more of the following ecosystem services: water retention, water quality improvement (e.g., nitrate, phosphate, sediment, or pesticide retention), wetland habitat (for birds, aquatic organisms, and pollinators), crop yield improvement, and other benefits (e.g., recreation, education, aesthetic appreciation, greenhouse gas retention). We found the five edge-of-field practices were all effective at removing nitrate with varying degrees of other potential benefits and disservices (e.g., greenhouse gas production). Drainage water recycling and water quality wetlands have the potential to provide the most co-benefits as they provide surface water systems for capturing surface flows in addition to tile drainage while also potentially providing habitat and recreation opportunities. However, the following research needs are identified: 1) the disservices and benefits associated with drainage water recycling have not been adequately evaluated; 2) surface flow dynamics are understudied across all reviewed management practices; 3) a complete accounting of phosphorus species and flow pathways for all management practices is needed; 4) field evaluations of the habitat benefit of all management practices are needed; and 5) greenhouse gas dynamics are understudied across all management practices. While all management practices are expected to reduce nitrate loads, addressing these knowledge gaps will help inform holistic management decisions for diverse stakeholders across the US Corn Belt.

1. Introduction

Agricultural tile drainage systems remove excess water from fields and are a major reason the United States (US) Corn Belt and other regions around the world have such productive cropland (Evans and Fausey, 2015; Fausey et al., 1995; Kumar et al., 2014; Mourtzinis et al., 2021). Unfortunately, tile drainage systems also expedite subsurface nitrate transport from agricultural systems, resulting in losses of nitrogen from the fields where it is needed and contributing to nitrate pollution in downstream surface waters (David et al., 2010; Goolsby et al., 2001; Schilling et al., 2020). Additionally, the capacity for tile drain systems to remove water from fields has led to the drainage of much of the historic wetland habitat important for amphibians and waterbirds in the US Corn Belt (Blann et al., 2009; Dahl, 1990). This multifaceted problem requires management solutions that can provide multiple co-benefits.

Edge-of-field management practices are needed in addition to in-field practices to address nutrient losses from agricultural systems and provide additional benefits. For example, in-field practices such as fertilizer application can be optimized with 4 R nutrient practices (i.e., right source, right rate, right time, and right place; Sharpley et al., 2019) while cover crops can bind or capture nutrients on the soil surface or in the soil column before they enter the tile drainage system (Liu et al., 2021; Mitsch et al., 2001; Thapa et al., 2018). However, even with optimized in-field management practices, losses of nutrients are still likely to occur in agricultural systems via subsurface tile drainage (Baker and Asae, 2003; Crumpton and Baker, 1993; Kalcic et al., 2018; Williams et al., 2018) and there is still a need for additional ecosystem service co-benefits such as habitat.

Established edge-of-field practices reduce tile flows moving downstream or create conditions that favor nitrate removal through denitrification and include: controlled drainage, denitrifying bioreactors, constructed water quality wetlands, drainage water recycling, and saturated buffers. While many studies have investigated and reviewed the water quality impacts associated with edge-of-field practices (e.g., Christianson et al., 2021; Crumpton et al., 2020; Hay et al., 2021; Helmers et al., 2022; Mulla, 2008; Sharpley et al., 2019; Singh et al., 2020; Vidon et al., 2019; Yuan et al., 2022), there is a lack of information and critical comparison regarding water quality benefits and other potential ecosystem service co-benefits (i.e., water retention, wetland habitat, crop yield improvement, recreational and educational opportunities, aesthetic appreciation, greenhouse gas retention) associated with their installation. Further, few studies have evaluated the potential negative impacts (e.g., greenhouse gas emissions), herein defined as disservices, of these management practices. The inclusion of these additional services and disservices provides a more holistic assessment that can inform management decisions and address a wider range of issues and stakeholder needs (Dale and Polasky, 2007; Swinton et al., 2007) while also limiting nitrate export.

We reviewed edge-of-field agricultural conservation practices for treating subsurface tile drainage with the goal of identifying a more complete understanding of the benefits, disservices, and research needs associated with their installation. Conservation practices appropriate for intercepting and potentially treating agricultural tile drainage identified and reviewed in this study include: controlled drainage, also known as water table management; drainage water recycling; denitrifying bioreactors; saturated buffers; and constructed or restored wetlands designed for water quality improvement, herein referred to as water quality wetlands. Studies that had information on one or more ecosystem services including water retention, water quality improvement, wetland habitat, crop yield improvement, and other benefits (e.g., recreation, education, aesthetic appreciation, greenhouse gas retention) were included in the review.

2. Methods

2.1. Literature review

The literature review was performed using the Web of Science database with the following keywords in different combinations: drainage water management, controlled drainage, constructed wetlands, wetland restoration, denitrifying bioreactors, saturated buffers, drainage water recycling, drainage water subirrigation, pollinators, water quality, pesticides, nutrients, sediment, habitat, wildlife, greenhouse gases, recreation, and education. See Supplemental Table 1 for all keywords and combinations employed in the literature search.

The review focused on studies performed in the Midwest United States. Agricultural conservation practices included in this evaluation were determined in an initial literature search, while relevant ecosystem services were identified by stakeholders across several meetings in 2018 and 2019 (Mitchell et al., 2022b) and literature review. Ecosystem services reviewed in this study are defined in Supplemental Table 2 using the National Ecosystem Services Classification System (NESCS Plus; Newcomer-Johnson et al., 2020).

We reviewed the identified studies to determine: 1) basic information about each conservation practice and how it works; 2) effects on water quality (i.e., nitrogen and phosphorus removal rates); 3) effects on greenhouse gases; 4) effects on other ecosystem service values; and 5) research gaps. Additional studies referenced by reviewed literature were also included where relevant.

2.2. Literature classification

Publications identified in the literature review were classified according to the conservation practice, location, the type of study (review, modeling, field study, laboratory study), and metrics evaluated (flow, crop yield, nitrate, total nitrogen, phosphorus, greenhouse gases, and other metrics). Controlled drainage studies involving subirrigation were classified as controlled drainage studies if a reservoir for collecting flows for recycling was not included. Studies where controlled drainage was paired with a reservoir for surface or subirrigation were classified in this review as drainage water recycling.

3. Results

The literature search obtained 119 relevant records relating to the identified agricultural conservation practices and one or more ecosystem services in the US Corn Belt (Table 1; Supplemental Table 3). Of these records we identified 58 studies on controlled drainage, 32 on water quality wetlands, 23 on denitrifying bioreactors, 14 on drainage water recycling, and 13 on saturated buffers (Table 1). Most records identified were focused solely on aspects of water quality (e.g., nitrate, phosphorus, sediment, and pesticide retention) or crop yield improvements, with just 10 studies on greenhouse gases and 9 studies on other potential benefits such as wetland habitat (for birds, aquatic organisms, and pollinators), recreation, education, and aesthetic appreciation (Table 1; Supplemental Table 3).

Table 1.

Summary of reviewed study counts by agricultural conservation practice, study type, and service or disservice metric. See Supplemental Table 3 for complete records.

Agricultural Conservation Practice (ACP)	Total Studies	Count of Flow	Count of Crop Yield	Count of Nitrate	Count of Total Nitrogen	Count of Phosphorus	Count of Greenhouse gases	Count of Other ES
Study type
Controlled Drainage (CD)	48	31	15	21	12	10	1	3
Field and Model	3	2		1	2
Field Study	23	13	9	11	1	4		2
Lab	1			1				1
Model	15	11	6	4	5	1	1
Review	6	5		4	4	5
Water Quality Wetland	20	1	1	18	5	10	3	6
(WQW)
Field and Model	1			1		1
Field Study	7			7	2	4		2
Model	4			2				1
Review	8	1	1	8	3	5	3	3
Denitrifying Bioreactor	18			11		6	3	2
(DB)
Field and Model	1						1
Field Study	7			4		1	1
Lab	5			4		4		2
Model	2			1
Review	3			2		1	1
Drainage Water Recycling	13	3	10	4	3	3		2
(DWR)
Field and Model	1		1
Field Study	6		5	1	1
Model	3	1	2	1		1
Review	3	2	2	2	2	2		2
Saturated Buffer (SB)	8			4		1	1
Field Study	5			4		1	1
Model	3
CD, WQW	5			3		2
Field and Model	1			1
Review	4			2		2
CD, WQW, DB	2	1	1	2		1	1	1
Field Study	1			1
Review	1	1	1	1		1	1	1
CD, WQW, DB, SB	1			1	1	1
Review	1			1	1	1
CD, WQW, SB	1		1	1
Model	1		1	1
CD, WQW, DB, SB, DWR	1					1
Review	1					1
WQW, DB, SB	1			1	1	1	1
Review	1			1	1	1	1
WQW, SB	1		1
Review	1		1

3.1. Controlled drainage

Controlled drainage, also known as drainage water management (USDA-NRCS Code 554), is a technique that employs systems to manage the flow of subsurface water in the drainage system and therefore influence the water table height (Gilliam et al., 1979; Skaggs et al., 2012). Water control structures are installed on the tile lines at the edge of the field to seasonally interrupt tile drainage flow, allowing the user to control the field’s water table depth manually or automatically (Fig. 1). Water flowing through the tile drainage system reaches the control structures and increases in depth until it reaches the height of the water control gates where it can overflow and leave the field. In this way, the water table is raised or lowered, and the drainage of water and nutrients is controlled. Common practice is to raise the water table following harvest in the fall, not lowering it again until spring field preparation, planting, and crop establishment, after which it is raised again to potentially allow for subirrigation (Helmers et al., 2012, 2022; Skaggs et al., 2012). As the majority of tile drainage flows occur outside of the growing season in much of the Midwestern US and Canada (Helmers et al., 2022; Tan and Zhang, 2011), this can be an effective strategy for reducing overall tile drainage flows.

Fig. 1.

Controlled drainage design showing two water control structures used to raise the depth of the drainage outlet. Grey-blue color corresponds to groundwater. Numbers correspond to the following: (1) Tile drain outflow; (2) Surface runoff; (3) Flow moving downstream. Image adapted from trans formingdrainage.org.

The literature review identified 56 studies related to controlled drainage in the US Corn Belt Region, consisting of 12 review articles, 28 field studies, and 20 publications using a modeling approach including four studies that employed both field and modeling methods (Table 1; Supplemental Table 3). The literature was dominated by nitrate studies (28), while 17 studies pertained to crop yields, 13 related to total nitrogen, 14 related to forms of phosphorus, eight studies evaluated hydrologic flows, two studies related to greenhouse gases, and four studies evaluated or discussed other metrics including dissolved organic carbon (one study), herbicides (two studies), and wildlife habitat (one study; Table 1; Supplemental Table 3).

Studies in the Midwestern US have demonstrated that controlling the water table with controlled drainage can decrease the total water leaving the field via subsurface drainage during storm events by 22% (Lahdou et al., 2019) and on an annual basis by 8%–85% (Fig. 2; Helmers et al., 2022; Shedekar et al., 2021; Skaggs et al., 2012). Drainage flow reductions are attributable to increased evaporation and seepage, or increased surface runoff associated with an elevated water table with more saturated conditions (Skaggs et al., 2010).

Fig. 2.

Potential conservation practice benefits based on literature review for several benefit categories. Each radial section corresponds to an approximate 20% improvement, with a maximum improvement shown of 40%. Sections marked with an asterisk (*) indicate uncertainty due to lack of study and or high dependency on practice management or local conditions. ¹Other Benefits and Wetland Habitat sections are qualitative comparisons while all other metrics are quantitative and represent the mean of evaluated studies. Other Benefits were sediment retention, recreation, education, aesthetic appreciation, pesticide retention, and greenhouse gas retention; these other benefits were not explicitly included as sectors of the wheel diagram because of limited information available about these ecosystem services. CD= Controlled drainage; WQW= Water quality wetland; DB = Denitrifying bioreactor; DWR = Drainage water recycling; SB= Saturated buffer.

Total drainage flow reductions result in reduced downstream nutrient loading, particularly nitrogen, with the reviewed studies in the US Midwest reporting a mean reduction in nitrate losses of 45% across all identified studies with a range of study means of 20%–99%, (Table 2; Supplemental Table 3; Helmers et al., 2012; Helmers et al., 2022; King et al., 2022; Shedekar et al., 2021; Skaggs et al., 2012), a standard deviation of 6.5, and an average area-normalized reduction rate of 13.7 kg N ha⁻¹ yr⁻¹ (Std. Dev. = 19.6). These US Midwest reduction rates are comparable to rates from a global meta-analysis that identified an average nitrate removal efficiency of 50% with a range of 19%–82% and an area normalized removal rate of 12 kg N ha⁻¹ yr⁻¹ (Carstensen et al., 2020). Similarly, Ross et al. (2016) identified a global mean nitrate-N removal efficiency of 48% and an area-normalized removal rate of 11.6 kg N ha⁻¹ yr⁻¹. However, many studies identified in our review did not assess or report nitrate or total nitrogen loading in surface runoff, flow pathways that can potentially increase due to controlled drainage and therefore offset some of the reductions in loading via tile drains (Ale et al., 2012; Ross et al., 2016).

Table 2.

Benefit comparison (mean/range of study means where quantified) of edge-of-field conservation practices for reducing nutrient export from agricultural tile drainage systems. N = Number of studies included in estimate for each metric. See Supplemental Table 3 for complete records, including total nitrogen and phosphorus species. CD= Controlled drainage; DB = Denitrifying bioreactor; SB=Saturated buffer; WQW= Water quality wetland; DWR = Drainage water recycling.

Edge-of-Field Practice	Treats Surface Runoff as well as Tile Drainage?	Combinable with Other Edge-of-Field Practice?	Scale of Operation	Nitrate-Nitrogen Removal Efficiency (%)	Total Phosphorus Removal (%)	Impact on Crop Yields?	Potential Reduction in Downstream Flows?	Potential Habitat Benefits?	Additional Ecosystem Service Value?
Controlled Drainage	No	Yes (DB, SB)	Field	Mean: 45% Range: 20%–99% N: 14	Mean: 13% Range: NA N: 1	Minimal to Positive	Minimal to Yes	Minimal to Yes	Miminal
Denitrifying Bioreactor	No	Yes (CD,SB, WQW)	Field	Mean: 42% Range: 20%–63% N: 7	Amendments Needed	Minimal	Minimal	Minimal	Minimal
Saturated Buffer	No	Yes (CD,DB,WQW)	Field	Mean: 37% Range: 8%– 84%a N: 1	Unknown	Minimal	Minimal	Minimal	Minimal
Water Quality Wetland	Yes	Yes (DB,SB,DWR)	Field/Watershed	Mean: 47% Range: 33%–68% N: 8	Mean: 33 Range: 0%–76% N: 4	Minimal to Negative	Minimal	Yes	Yes
Drainage Water Recycling	Yes	Yes (DB,SB,WQW)	Field/Watershed	Mean: 31% Range: 24%– 37%a N: 1	Unknown	Positive	Minimal to Yes	Yes	Yes

^aRange reported is across multiple sites or model scenarios but from one publication.

There are potential water quality tradeoffs associated with controlled drainage as flows shift from subsurface to surface (Gilliam and Skaggs, 1986; King et al., 2022; Lahdou et al., 2019; Sadhukhan et al., 2019; Singh et al., 2007; Skaggs et al., 2010). For example, Shedekar et al. (2021) identified a 50% increase in surface runoff in controlled drainage treatments while a review by Ross et al. (2016) similarly identified surface runoff in controlled drainage treatments was 153% greater than free drainage treatments. Large increases in surface runoff could result in increased losses of dissolved reactive phosphorus, soil particles, and associated sorbed materials such as phosphorus or pesticides, but surface loading dynamics under controlled drainage are understudied (Ross et al., 2016). For example, the only identified study assessing surface flow pathways of phosphorus modeled a net 13% increase in total losses of phosphorus due to controlled drainage even though the model predicted an 85% reduction in phosphorus loading via tile drains (Table 2; Sadhukhan et al., 2019).

Another water quality disservice can occur if the anaerobic conditions generated in saturated soils with controlled drainage lead to pollutant swapping where nitrogen is removed but phosphorus in the soil is released; dynamics that appear to be driven by the availability of phosphorus in soils (King et al., 2015; Tan and Zhang, 2011). Nevertheless, a review of controlled drainage studies found an average dissolved reactive phosphorus reduction in tile drainage of 56% and an average total phosphorus reduction of 54%, also in tile drainage (Ross et al., 2016). However, many studies have not evaluated contributions of particulate phosphorus in tile-drainage or assessed and reported phosphorus in surface runoff; therefore more research is required before a complete picture is evident (King et al., 2015; Ross et al., 2016).

When wet periods precede drought during the growing season, controlled drainage can raise the water table and therefore conserve plant available water in the soil profile for growing crops (Skaggs et al., 2012) and result in increases in crop yields (Delbecq et al., 2012; Ghane et al., 2012) and greater crop nutrient uptake (Sunohara et al., 2014). However, yield improvements due to controlled drainage appear to be rare and site or year specific, as many studies suggested no change in yield and one study found reduced yields of corn (Table 2; Helmers et al., 2012; Skaggs et al., 2012). A synthesis of corn yield studies in the Midwestern US and North Carolina by Youssef et al. (2023) suggests small increases in corn yields may result from controlled drainage in mild to moderate drought conditions while small corn yield decreases may occur with controlled drainage in wet conditions.

It remains unclear how controlled drainage influences greenhouse gas emissions. Increased soil saturation due to controlled drainage likely leads to increases in denitrification but could result in incomplete denitrification or methanogenesis and therefore increase the production and release of nitrous oxide or methane, respectively (Hagedorn et al., 2022; Nangia et al., 2013). No studies that we are aware of have measured the greenhouse gas implications of controlled drainage in the US Corn Belt region. However, Hagedorn et al. (2022) in Maryland, Nangia et al. (2013) in Eastern Ontario, Canada, and Kliewer and Gilliam (1995) in North Carolina found no significant effects of controlled drainage on greenhouse gas emissions, while Wang et al. (2016) modeled nitrous oxide emissions in Iowa under future climate scenarios and predicted an increase in nitrous oxide emissions of 7.5% due to controlled drainage. Greenhouse gas production and release are likely to be highly variable from site to site and with controlled drainage management decisions, and therefore must be studied throughout the Corn Belt under varying conditions moving forward.

There are potential aquatic resources provided for migrating and year-round waterbirds with controlled drainage if the water table is raised to a height that floods portions of the field during late winter or early spring (Fig. 2; Table 2). Common practice is to raise the water table within 25–50 cm of the soil surface (Helmers et al., 2022; Skaggs et al., 2012), resulting in inundation of any low-lying or depressional areas for prolonged periods. However, there were no identified studies that mentioned this potential benefit. Therefore, the feasibility of maintaining a high enough water table, the tradeoffs associated with this strategy, and barriers to implementation need to be investigated through stakeholder engagement, modeling, and field studies.

In summary, controlled drainage improves nitrate retention and potentially water retention, while also potentially resulting in small increases in crop production and seasonal wetland habitat in the early spring if designed and managed accordingly (Fig. 2; Table 2). However, there is the potential for pollutant swapping where phosphorus is lost due to increases in surface runoff and/or leaching of bound phosphorus due to anaerobic conditions. These potential benefits and tradeoffs will depend on the local conditions and system management, and nearly all aspects of controlled drainage require additional study to further elucidate these patterns.

3.2. Water quality wetlands

Water quality wetlands are restored or constructed wetlands (USDA-NRCS Code 656) that are designed to intercept tile drainage systems from multiple farm properties while also providing habitat (Fig. 4; Crumpton, 2001; Day, 2003; Kadlec, 2012; Mitchell et al., 2022b; Mitsch et al., 2001; Zedler, 2003). As these systems are designed to operate at full pool (Mitchell et al., 2022b) and have low seepage rates (Crumpton et al., 2020), they are expected to have minimal effects on total flows and downstream flooding but have the potential to address nitrate losses from tile and surface runoff due to conditions that favor denitrification.

Fig. 4.

Denitrifying bioreactor layout showing the use of water control structures to divert a portion of tile-drainage into a woodchip bioreactor. Large arrow indicates direction of subsurface flow through the woodchip bioreactor. Grey-blue color corresponds to groundwater. Numbers correspond to the following: (1) Tile drain outflow; (2) Bypass flow when bioreactor is full; (3) Flow moving downstream.

In total the literature review identified 31 publications in the Corn Belt that pertained to water quality wetlands, 11 of which also discuss at least one other conservation practice (Table 1). Sixteen of the 31 publications were review articles, ten studies were field studies (two of these also included modeling), and seven studies used a modeling approach (Table 1; Supplemental Table 3). Twenty-three studies pertained to nitrate, 15 related to some form of phosphorus, seven related to total nitrogen, five studies related to greenhouse gases, three studies pertained to crop yields, two studies evaluated hydrologic flows, and seven studies evaluated or discussed other metrics including wildlife habitat (four studies), pesticides (two studies), and carbon (one study; Table 1). However, excluding review articles, there were only 11 nitrate studies, two total nitrogen studies, five phosphorus studies, one carbon, one pesticide, and one wildlife study (Table 1).

Recent studies of water quality wetlands in the Midwest suggest a large capacity for nitrate reduction with a mean of 47%, a standard deviation of 15.7, and a study mean range of 35%–86% (Table 2; Supplemental Table 3; Crumpton et al., 2020; Kovacic et al., 2000; Kovacic et al., 2006; Mitsch et al., 2005a; O’Geen et al., 2010). Similarly, a global meta-analysis by Carstensen et al. (2020) found an average nitrate removal efficiency of 41% with a range from −8% to 63% and a wetland area-normalized rate of 600 kg N ha⁻¹ yr⁻¹, while a global review by Messer et al. (2021) identified mean nitrate concentration reductions ranging from 14% to 45%. For comparison, water quality wetlands in Iowa removed a wetland area-normalized average of 1500 kg nitrate-N ha⁻¹ yr⁻¹ (Crumpton et al., 2020).

Water quality wetlands can reduce total phosphorus loading, although retention is highly variable (Carstensen et al., 2020; Land et al., 2016; Messer et al., 2021) and the ability for the wetland to operate as a sustainable sink for phosphorus is dependent on phosphorus accretion in wetland sediments (Anderson et al., 2005; Mitsch et al., 2005b) or the removal of biomass (Gordon et al., 2021; Lenhart et al., 2016). Identified studies in the Midwest reduced total phosphorus losses on average by 33% with a study mean range of 0%–76% and standard deviation of 32.8 (Table 2). Global metanalyses by Land et al. (2016) and Carstensen et al. (2020) identified phosphorus removal rates of 12 kg phosphorus ha⁻¹ yr⁻¹ and 6.8 kg phosphorus ha⁻¹ yr⁻¹, respectively, with Carstensen et al. (2020) identifying a net phosphorus retention efficiency of 18%.

Water quality wetlands and drainage water recycling are the only two management practices highlighted here that can, if designed appropriately, address subsurface and surface water quality directly (Table 2). Accordingly, water quality wetlands have the potential to reduce flows of sediments downstream as well as any phosphorus or pesticides that are adhered to them, but these dynamics are understudied.

Water quality wetlands also can potentially provide wetland habitat for a broad range of aquatic organisms and waterfowl (Ballard and Jones, 2021; Fennessy and Craft, 2011; Knutson et al., 2004; Mitchell et al., 2022a; Reeves et al., 2016; Swanson et al., 2019), and if installed with a surrounding habitat buffer area as is typical for this practice, these systems can likely also provide grassland bird and pollinator habitat (Mitchell et al., 2022a).

If installed on productive cropland, water quality wetlands can reduce crop production due to land replacement (Table 2; Mitchell et al., 2022b), and like other management practices that rely on anaerobic conditions for denitrification there is potential for the production and release of nitrous oxide and methane. Across a limited number of studies, it appears that water quality wetlands are likely to release nitrous oxide at an equivalent rate to fertilized cropland (Mitchell et al., 2022b) but that methane release can be substantial and elevated compared to the cropland that the wetland is replacing (Gleason et al., 2009; Mitchell et al., 2022b; Mitsch et al., 2013). Water quality wetlands can, however, sequester carbon in organic matter for centuries, potentially offsetting this negative effect (Euliss et al., 2006; Fennessy and Craft, 2011; Mitchell et al., 2022b; Mitsch et al., 2013).

Water quality wetlands can provide substantial reductions in nitrate loads, smaller reductions in phosphorus, increased wetland and grassland habitat, and potentially provide other benefits like sediment retention, recreation, and education benefits (Fig. 2; Table 2). These benefits appear to outweigh the potential disservices of small reductions in crop production. However, water quality wetlands are expensive to install and often require specific site requirements that narrow potential locations. More research is required to understand their net impacts on methane and other greenhouse gases. Additionally, emerging types of water quality wetlands like reconstructed oxbows, a reexcavated isolated river meander that can receive tile drainage and remove nitrate-nitrogen (Schilling et al., 2017), require more study to understand the services and disservices they provide.

3.3. Denitrifying bioreactors

Denitrifying bioreactors (USDA-NRCS Code 605) are designed to remove nitrate via denitrification. Using water control structures, a portion of tile-drainage water is directed through a buried pile of organic material, commonly woodchips, which provide a carbon source and substrate for denitrifying microbes (Fig. 3; Hartfiel et al., 2022). Additional amendments can also be added to the woodchips to bind phosphorus, pesticides, and other pollutants. After moving slowly through the woodchips, drainage water returns to the drainage system where it can be routed to surface water or other management practices. Once the water storage capacity of the reactor is reached, any additional tile-flow exceeding the outflow rate from the bioreactor is routed to bypass the bioreactor. Bioreactors installed in the Midwest are commonly designed to treat 3–20 ha of agricultural land (Christianson et al., 2021).

Fig. 3.

Water quality wetland layout. Grey-blue color corresponds to groundwater. Numbers correspond to the following: (1) Tile drain outflow; (2) Surface runoff; (3) Flow moving downstream.

The literature review identified 23 studies that pertained to denitrifying bioreactors, with five of these also investigating at least one other practice (Table 1). Eight of these studies were field studies, five were at the lab scale, and six studies consisted of review articles. Only two studies used a modeling approach. Fifteen studies focused on nitrate, ten evaluated phosphorus, three studies investigated greenhouse gases, one study investigated bacterial removal, and one study evaluated pesticide removal (Table 1; Supplemental Table 3).

Denitrifying bioreactors are effective at removing nitrate, with field studies in the US Midwest reporting a mean nitrate removal capacity of 42% with a study mean range of 20%–63%, standard deviation of 11.6, and area normalized rates ranging from 0.5 to 29 kg N ha⁻¹ (Table 2; Supplemental Table 3; Christianson et al., 2012; Christianson et al., 2021; Jaynes et al., 2008). A global meta-analysis identified an average nitrate removal efficiency of 40% with a range of 6%–79% and an area-normalized reduction rate of 5940 kg N ha⁻¹ yr⁻¹ (Carstensen et al., 2020). However, the nitrogen removal capacity of bioreactors is limited by the amount of water that can be routed through the reactor itself, as well as the expected lifespan of the woodchips used to fuel denitrification (Christianson et al., 2021; Ghane et al., 2018; Hartfiel et al., 2022; Moorman et al., 2010; Schaefer et al., 2021). Additionally, many studies did not evaluate or report nitrate loading reductions or were too short to quantify an annual nitrate reduction due to denitrifying bioreactors.

As with the other denitrification-dominated practices, there is potential for anaerobic conditions generated in denitrifying bioreactors to result in pollutant swapping. For example, studies have indicated the release of phosphorus from the woodchip media in denitrifying bioreactors for several months after installation (Bell et al., 2015; Christianson et al., 2021; Husk et al., 2018). However, bioreactors can operate as phosphorus sinks following this initial leaching period, especially if materials capable of sorbing phosphorus such as steel slag, iron turnings, fly ash, or activated alumina are incorporated into the woodchip media or as a two-stage filter (Christianson et al., 2021; Goodwin et al., 2015; Hua et al., 2016; Husk et al., 2018; Li et al., 2018). Similarly, Hassanpour et al. (2019) found that the addition of biochar resulted in the removal of 90% of atrazine inputs in a lab trial, while Ilhan et al. (2011) demonstrated potential sorption of pesticides including atrazine in woodchip-based laboratory bioreactors without amendments. However, the capacity to retain potential pollutants such as phosphorus and atrazine is likely not sustainable, as fresh materials must be added to recharge its capacity. Denitrifying bioreactors may also result in the reduction of sulfate and production of methylmercury if nitrate becomes limiting, but this is an area requiring additional study (Bell et al., 2015; Hartfiel et al., 2022). Beyond removal of nitrate and potential retention of other pollutants and pathogenic bacteria (Soupir et al., 2018), denitrifying bioreactors have limited capacity to provide additional benefits (Fig. 2; Table 2).

As with controlled drainage, saturated buffers, and water quality wetlands, anaerobic conditions in biorectors promote nitrate-nitrogen loss via denitrification, but can also promote the production of nitrous oxide and methane (Greenan et al., 2009; Hartfiel et al., 2022). Bioreactors can be sources of carbon dioxide, nitrous oxide, and methane (David et al., 2016; Davis et al., 2019b; Elgood et al., 2010; Ghane et al., 2015; reviewed in Hartfiel et al., 2022) but it is important to consider these emissions relative to emissions from the land that the bioreactor is replacing and the emissions from the decomposition of woodchip media that would be released regardless of the creation of the denitrifying bioreactor. For example, Elgood et al. (2010) in Southern Ontario, Canada, Bock et al. (2018) in Virginia, and Moorman et al. (2010) in Iowa found nitrous oxide emissions were similar to reported emissions from cropland while Ghane et al. (2015) in Ohio and Greenan et al. (2009) in laboratory columns found complete denitrification with low nitrous oxide emissions was common. In a study of a nine year old denitrifying bioreactor in Iowa, Moorman et al. (2010) found total nitrous oxide emissions and dissolved nitrous oxide levels in effluent from the bioreactor were not significantly different from an untreated control. However, few other studies have quantified methane emissions or dissolved nitrous oxide emissions in treated water from bioreactors or considered downstream losses of nitrous oxide (Christianson et al., 2021; Davis et al., 2019b); therefore greenhouse gas dynamics require additional study.

In summary, denitrifying bioreactors are a powerful tool for addressing nitrate losses from agricultural systems with very few expected disservices. However, they provide almost no additional benefits beyond nitrate retention (Fig. 2), while also requiring relatively frequent maintenance to replace woodchips to fuel denitrification.

3.4. Drainage water recycling

Unlike the other practices highlighted in this work, drainage water recycling (USDA-NRCS Code 447) does not primarily rely on denitrification for nitrogen removal. Instead, drainage water recycling combines a reservoir with a pump and irrigation system so that drainage water can be collected and reapplied for irrigation when needed (Fig. 5; Baker et al., 2012; Hay et al., 2021). Irrigation can be applied via surface application or pumped back into the tile system as subirrigation if tile spacing and soil conditions are appropriate (Hay et al., 2021). For this review, studies incorporating an irrigation pond with water derived primarily from other sources than drainage (e.g., municipal water) were not considered as drainage water recycling practices.

Fig. 5.

Drainage water recycling layout. Grey-blue color corresponds to groundwater. Numbers correspond to the following: (1) Tile drain outflow; (2) Potential surface runoff; (3) Drainage water reuse for irrigation. Image adapted from transformingdrainage.org.

The literature review identified 14 total studies involving drainage water recycling in the Midwest, consisting of seven studies using field evaluations (including one field and modeling study), four using a modeling approach, and four studies that were reviews (Table 1). Studies were overall focused on crop yields (10 studies), water quality (four nitrate studies, four phosphorus studies, three total nitrogen studies), and hydrologic flows (three studies), with only two studies of wildlife habitat (Table 1; Supplemental Table 3).

Capturing and storing drainage water using drainage water recycling theoretically reduces the flow of water and nutrients downstream (Fig. 2; Table 2), while pumping this water onto fields when needed recharges the ability of the reservoir to capture drainage (and surface runoff) and reduces external inputs of water for irrigation (Hay et al., 2021). No field studies in the US Midwest were identified that directly measured nutrient losses in drainage water recycling systems alone (i.e., no adjacent water quality wetland), defined as having a reservoir that captures tile drainage used for irrigation. However, Reinhart et al. (2019) modeled drainage water recycling systems in Iowa and Indiana and predicted nitrogen loss reductions due to drainage water recycling ranging from 24% to 37% and soluble reactive phosphorus loss reductions ranging from 21% to 39%, amounting to area-normalized removal rates of 9–11 kg nitrate-nitrogen ha⁻¹ yr⁻¹ and 0.02–0.05 kg phosphorus ha⁻¹ yr⁻¹ (Fig. 2; Table 2; Supplemental Table 3). Adding or supplementing irrigation also has the potential to increase crop yields by 4%–61% for corn and 4%–31% for soybeans (Fig. 2; Table 2; Supplemental Table 3; Allred et al., 2014a; Baule et al., 2017; Hay et al., 2021; Kaur et al., 2021) and reduce year-to-year variability in crop yields (Willison et al., 2021).

Drainage water recycling reservoirs can potentially provide aquatic habitat benefits, particularly if there is an additional wetland portion adjacent to and connected to the main reservoir (Fig. 2; Table 2; e.g., Brown et al., 1998). Smiley and Allred (2011) found both amphibians and fish in drainage water recycling reservoirs and adjacent constructed wetlands in Northwest Ohio. However, aquatic habitat benefits will depend on how the system is managed and where it is used, and the Smiley and Allred (2011) study was the only study identified that measured or modeled habitat benefits in these systems. Partially draining the drainage water recycling reservoir seasonally for irrigation needs may have negative consequences for certain wetland-dependent organisms but may benefit other organisms like shorebirds that can extract food from exposed, recently flooded areas. An adjacent wetland pool may help control for these fluctuations, but also will necessitate more land out of production and/or more expense to create and may also harbor fish sourced from the reservoir which may impair amphibian habitat in the adjacent wetland (Smiley and Allred, 2011).

There are potential disservices with drainage water recycling. For example, adding a reservoir can replace productive cropland and may make drainage water recycling financially feasible only in certain locations; although a changing climate that favors higher precipitation in the spring and less precipitation in the growing season (USGCRP et al., 2018) will certainly improve the economic outlook for this practice (Baule et al., 2017; Hay et al., 2021). There are also potential increases in greenhouse gases associated with the installation of a reservoir, but no studies were identified that investigated this aspect. As with other practices, the net greenhouse gas impacts will greatly depend on the greenhouse gas emissions from the land that the reservoir is replacing and how the water level in the reservoir is managed. Lastly, salts can become concentrated due to irrigation water recycling (Hay et al., 2021) while E. coli concentrations can assume high concentrations during warmer months (Haverstock et al., 2017) in drainage water recycling reservoirs, both aspects to be considered during installation and management.

Overall, the emerging practice of drainage water recycling has the potential to increase crop yields, water retention, nitrate retention, and phosphorus retention while supplying aquatic habitat (Fig. 2), but all these aspects require further study. Other benefits such as sediment capture may be possible if the reservoir is located and designed to capture surface runoff as well as tile drainage. Incorporating adjacent constructed wetlands that drainage water can be routed through before entering the reservoir for irrigation may be beneficial for further reducing losses of nitrogen (Allred et al., 2014b) and other potential contaminants (Haverstock et al., 2017) while potentially providing additional habitat. Other potential benefits to the landowner associated with drainage water recycling include recreational benefits like hunting, fishing, and potential improvements in aesthetics due to the reservoir (Fig. 2). However, drainage water recycling systems are typically very expensive (Sellars et al., 2022), and as of now are likely to make economic sense only in areas where irrigation is absolutely required; although shifting weather patterns may cause these economic dynamics to shift in favor of this practice for much of the US Midwest.

3.5. Saturated buffers

Like controlled drainage and denitrifying bioreactors, saturated buffers (USDA-NRCS Code 604) use water control structures installed along drainage tile outlets. In the case of saturated buffers, these control structures are used to divert and disperse a portion of the subsurface drainage into a saturated riparian area (Fig. 6; Jaynes and Isenhart, 2014). Saturated buffers are primarily designed to remove nitrate-nitrogen via the process of denitrification and are comparatively inexpensive to install but rely on existing soil banks that have relatively high levels of organic matter to fuel denitrification (Chandrasoma et al., 2019).

Fig. 6.

Saturated buffer layout. Grey-blue color corresponds to groundwater. Numbers correspond to the following: (1) Tile drain outflow; (2) Drainage water flow through saturated buffer; (3) Flow moving downstream. Image adapted from transformingdrainage.org.

The literature review identified 13 total studies pertaining to saturated buffers in the Corn Belt Region. Five of these studies were field studies, four used modeling approaches, and four were review articles. Work was dominated by nitrate studies with seven articles, followed by four articles involving forms of phosphorus, two articles involving crop yields, two review articles pertaining to total nitrogen, and two articles involving greenhouse gas emissions (Table 1; Supplemental Table 3).

As saturated buffers become saturated, a large portion of the nitrate moving through the buffer is converted through denitrification to dinitrogen gas, therefore reducing nitrate in the treated water leaving the buffer (Groh et al., 2019). Nitrate retention using saturated buffers was improved by a mean of 37% with a range of 8%–84% across 17 total site-years in six sites located throughout Iowa (Table 2; Jaynes and Isenhart, 2019). A recent global meta-analysis by Carstensen et al. (2020) reported an average nitrate retention rate of 37% when drainage water not captured by a saturated buffer was included, while 68% of nitrate was removed in the water reaching the saturated buffer, altogether amounting to an area-normalized retention rate of 230 kg N ha⁻¹ yr⁻¹. As with other conservation practices highlighted in this manuscript that rely on denitrification, there is potential for anaerobic conditions generated in saturated buffers to result in pollutant swapping where nitrogen is removed but phosphorus in the soil is released (King et al., 2015; Vidon et al., 2019).

While small, there are potential habitat benefits provided by these systems which are commonly planted with perennial vegetation (Groh et al., 2019), although habitat was not formally evaluated in any of the identified studies. Saturated buffers could be planted with perennial bioenergy crops like switchgrass to also provide an economic benefit to the landowner (Kreig et al., 2019).

There are few additional benefits provided by saturated buffers beyond nitrate retention (Fig. 2; Table 2), and there is uncertainty regarding greenhouse gas emissions. Like other systems that rely on saturated conditions to promote denitrification to remove nitrate-nitrogen, saturated buffers may result in higher nitrous oxide emissions. This is of particular concern with saturated buffers that commonly involve non-optimized conditions for denitrification (Jaynes and Isenhart, 2019). However, a study of two saturated buffers in Iowa suggests that overall nitrous oxide emissions, including downstream emissions due to in-stream denitrification, may be similar to other riparian systems and lower than fertilized corn and soybean systems (Davis et al., 2019a). But with limited work beyond the Davis et al. (2019a) study, and no identified studies evaluating methane or carbon dioxide fluxes, greenhouse gas emissions from saturated buffers require additional study.

In summary, saturated buffers can be very effective for reducing nitrate export from agricultural systems but are unlikely to provide many other substantial benefits (Fig. 2). However, the anticipated disservices associated with this practice and the expense of installing saturated buffers in appropriate locations are minimal, making saturated buffers a targeted solution for addressing nitrate losses.

4. Discussion

All identified conservation practices for handling agricultural subsurface tile drainage were found to be effective for reducing nitrate moving downstream, however their relative effectiveness and ability to provide additional benefits varied substantially (Table 2; Fig. 2). The identified conservation practices also varied with regards to their scale of operation (field vs watershed scale), with drainage water management and water quality wetlands identified as having the capacity to treat subsurface drainage from larger areas at the small watershed scale. These watershed-scale approaches also can intercept and treat surface runoff if designed and situated appropriately (Table 2), an aspect that is missing from most field-scale approaches.

4.1. Water balance

To appreciably reduce water flowing downstream, a conservation practice must either have a very large storage capacity or the ability to recharge this capacity relatively quickly. Of the practices considered here, controlled drainage and drainage water recycling have the potential to affect soil-water balances in crop fields. However, these effects will depend on local conditions as well as the quantity and timing of water availability and require additional study (Fig. 2). Although drainage water recycling and water quality wetlands have the potential to intercept and alter the dynamics of surface runoff, it is unlikely that any of the practices considered can substantially alleviate downstream flooding during major flow events (Fig. 2; Table 2).

4.2. Nitrate retention

On an area basis, water quality wetlands provided by far the highest nitrate reduction in studies performed in the Midwest. However, the global meta-analysis by Carstensen et al. (2020) suggests that denitrification bioreactors may have similar or even greater nitrate reduction potential as water quality wetlands, but reduction rates are highly dependent on nitrate inputs and hydraulic loading rates and there are also inconsistencies in statistical reporting and what area is used to normalize reductions.

On an efficiency basis, all practices except for drainage water recycling provided potential nitrate reductions in a global analysis ranging up to or exceeding 60% (Table 2). However, to our knowledge, no published studies have directly evaluated drainage water recycling nitrate dynamics and there is only one modeling study, where drainage water recycling was found to reduce nitrate losses by up to 37% (Reinhart et al., 2019). Additional study of this practice under different conditions and locations is therefore required.

4.3. Phosphorus retention

Overall, phosphorus dynamics associated with tile-drainage management practices have received much less study than nitrate (36 studies compared to 66 studies of nitrate) and are complicated by the various species and transport pathways of phosphorus in agricultural systems. What is evident is that phosphorus dynamics associated with the reviewed conservation practices are at least partially driven by the availability of phosphorus within the soil or media of the conservation system relative to levels of phosphorus in the incoming untreated drainage water. Therefore, the effects of management practices on phosphorus dynamics may vary widely with conservation practice design, operation, and location (Fig. 2; Table 2).

As with nitrate, water quality wetlands appear to have the highest capacity for reducing phosphorus export (Table 2), but many studies have also shown phosphorus export from these systems. The ability of water quality wetlands and drainage water recycling practices to potentially capture surface runoff and sediments may also give them an advantage in this area. Conversely, using controlled drainage to reduce flows of drainage water may increase surface runoff overall, potentially leading to elevated losses of phosphorus via runoff (Table 2), but phosphorus dynamics in these and other systems require additional study. The use of amendments for phosphorus sorption in denitrifying bioreactors also requires further study, as these amendments may prove to be beneficial at least in the near term for phosphorus removal (Singh et al., 2020).

4.4. Pesticide retention

As pesticides are frequently sorbed to soil particles and transported in surface runoff, practices that can capture surface runoff and associated sediments can be effective for capturing and retaining pesticides over the short term. In high energy systems, however, this retention will not be sustainable unless sediments are routinely removed from the wetland or reservoir or pesticides degrade quickly in this environment. Practices like controlled drainage that potentially increase surface runoff, on the other hand, may increase the loss of pesticides via surface runoff and erosion even though they may decrease losses of pesticides in tile drainage.

Denitrifying bioreactors may also decrease pesticide loading. Ilhan et al. (2011) demonstrated the potential for woodchip bioreactors to sorb pesticides including atrazine without amendments, while Hassanpour et al. (2019) found almost complete removal of atrazine inputs in a bioreactor amended with biochar. However, the pesticide dynamics of bioreactors and other practices will depend on pesticide characteristics, application technique, and the timing of pesticide applications as well as local site conditions.

Unfortunately, only five studies pertaining to pesticides and the reviewed edge-of-field practices were identified, with two focused on controlled drainage, two focused on water quality wetlands, and one focused on denitrifying bioreactors. Therefore, pesticide dynamics require additional study under varying conditions throughout the Corn Belt.

4.5. Crop production

On the positive side, drainage water recycling has the greatest potential for improving crop yields and reducing year-to-year yield variability (Fig. 2; Table 2), although some of these benefits may be offset by the necessary installation of a reservoir if it is installed on and replaces productive land. Controlled drainage may also provide some crop production benefits due to subirrigation, but these benefits are not consistent and are expected to be highly variable with location, drainage system design, and climate (Table 2).

Water quality wetlands, on the other hand, generally have larger footprints and can result in areas taken out of production (Table 2), but are ideally sited for installation in lower productivity soils to minimize this tradeoff (Mitchell et al., 2022b). Practices with a smaller footprint that can be installed in poor soils directly adjacent to streams or drainage ditches, such as denitrifying bioreactors and saturated buffers, are likely to have minimal impacts on crop production as very little productive land must be replaced (Table 2).

4.6. Habitat

Habitat related benefits associated with the reviewed conservation practices are vastly understudied. Water quality wetlands and drainage water recycling both have the potential to provide aquatic habitat (Fig. 2; Table 2), with the potential for additional grassland and/or pollinator habitat if they are installed with a surrounding vegetated buffer planted with perennial vegetation (Table 2). Controlled drainage can also provide ephemeral aquatic habitat if managed to allow low-lying areas such as drained pothole wetlands to flood seasonally (Table 2; Fig. 2). Other practices may provide small portions of perennial vegetation for birds and pollinators, but these benefits are likely to be minimal (Table 2).

4.7. Greenhouse gases

Like habitat, very few studies have evaluated the greenhouse gas implications and tradeoffs from conservation practices designed to reduce tile-drainage nutrient runoff. Full evaluation of these implications and tradeoffs requires an inventory of the fluxes of nitrous oxide, methane, and carbon dioxide and sequestration of carbon in soils, sediments, and plant material.

In general, many of the identified conservation practices in this review are designed to optimize denitrification, with frequently similar or lower nitrous oxide emissions relative to the land they are replacing. Optimized denitrification management practices also reduce nitrate moving downstream, potentially reducing overall nitrous oxide emissions on a landscape scale if less efficient denitrification occurs under suboptimal conditions downstream. However, very few studies have evaluated the nitrous oxide emissions from any practice, and even fewer have directly measured dissolved nitrous oxide in tile-drainage and treated effluent. Therefore, there is widespread uncertainty in quantifying net nitrous oxide emissions from these systems.

Similar to nitrous oxide emissions, generating anaerobic conditions for denitrification can also result in methane emissions, but these have not been adequately characterized in any of the identified conservation practices. Additionally, it is important to consider methane emissions relative to other forms of carbon, as methane may break down in the atmosphere well in advance of the release of sequestered carbon (Mitsch et al., 2013).

4.8. Other benefits and compatibility

Drainage water recycling and water quality wetlands have the potential to provide more additional benefits than other conservation practices as they provide often large surface water systems for surface runoff and sediment capture as well as recreational and educational benefits. Drainage water recycling and water quality wetlands can therefore potentially reduce downstream sedimentation and associated contaminants while also providing opportunities for hunting, sightseeing, and aesthetic landscape improvements (Fig. 2; Table 2). All systems highlighted in this document have the potential to provide at least limited educational benefits, but highly visible systems like drainage water recycling and water quality wetlands are likely to provide more opportunities than less visible systems like controlled drainage.

All the reviewed edge-of-field solutions in this manuscript can potentially be combined synergistically with in-field conservation practices and many of the edge-of-field practices as well. The only practical limitation is in combining controlled drainage with water quality wetlands or drainage water recycling, as these practices rely on full pools for denitrification or irrigation, respectively (Table 2).

5. Research needs

There remain several knowledge gaps that must be addressed, including.

1. Phosphorus: Across all reviewed practices, only five total studies focused in the Corn Belt were identified that reported total phosphorus export dynamics. Four of these five were related to water quality wetlands, illustrating the lack of research in this area. As the Great Lakes and many inland waters are particularly affected by phosphorus loading, it is essential that more research investigate the multitude phosphorus species and flow pathways for all relevant practices.

2. Surface runoff dynamics: Very few identified studies evaluated the surface runoff dynamics associated with conservation practice implementation. Sediments, pesticides, and phosphorus may be transported in surface runoff, and these fluxes may be lower or higher as a result of certain practices. It is therefore necessary to continue to study these dynamics in addition to subsurface flows, particularly related to the use of controlled drainage.

3. Habitat: Wetland and grassland habitat are lacking throughout the Corn Belt. Edge-of-field practices that can provide habitat or resources in addition to addressing water quality are needed. However, only seven studies were identified that discussed this important service, all but one consisting of review articles. In particular, the habitat benefits of drainage water recycling and the depressional ponding dynamics associated with controlled drainage management require further study.

4. Greenhouse gases: Edge-of-field conservation practices may influence the release of the potent greenhouse gases nitrous oxide and methane. However, only ten studies were identified relating to greenhouse gas emissions, six of which were review articles. Greenhouse gas fluxes therefore require additional study in the reviewed conservation practices, particularly as they relate to: carbon sequestration in these same systems; the emissions from the land prior to practice implementation; and fluxes of nitrous oxide downstream. In addition, dissolved nitrous oxide fluxes in tile drainage and conservation practice outflows must be quantified to ensure all meaningful fluxes are captured in these studies.

5. Drainage water recycling and saturated buffer practices: These emerging conservation practices have not been adequately evaluated, particularly with full-scale field studies. While saturated buffers have received some attention, almost all of the field studies in the Corn Belt have taken place in Iowa and therefore this practice needs to be evaluated in different locations and under different conditions.

6. Conclusions

The evaluated edge-of-field and watershed scale conservation practices for treating tile drainage were all found to be highly effective for reducing nitrate loading with varying degrees of other potential benefits and disservices. Overall, subsurface flows and contaminants, particularly nitrate and dissolved phosphorus, were relatively well-documented through field, lab, modeling, and review studies. Other flow pathways, forms of nutrients, and contaminants such as pesticides received much less study, while other services such as habitat and disservices such as greenhouse gas emissions received very little attention overall.

Controlled drainage received the most attention in the literature and was found to reduce nitrate loading via subsurface drainage. However, disservices associated with a seasonally high water table, including potential emissions of greenhouse gases and the potential for increased surface runoff and surface transported contaminants, require further study.

Water quality wetlands were also effective for reducing subsurface nitrate while also potentially reducing surface transported contaminant flows. However, surface flows and surface-transported contaminant inputs and outputs from water quality wetlands were rarely quantified, while greenhouse gas emissions also require additional study. Water quality wetlands can provide habitat and resources for flora and fauna, and recreational and educational opportunities for humans, but these aspects have rarely been quantified.

Denitrifying bioreactors were found to reduce subsurface nitrate flows with very few documented additional services and disservices. They can however be amended with materials to bind other contaminants including phosphorus and pesticides.

Saturated buffers and drainage water recycling overall received very little study. Almost all of the published field trials for saturated buffers took place in Iowa, while no field studies evaluating services besides crop yields were identified where drainage water recycling was used on its own without other sources of water or other treatment technologies. Both practices offer the potential for large reductions in nitrate, while drainage water recycling may provide other benefits including crop yield improvements, habitat, and recreation. Both practices also require further study to understand their net impact on greenhouse gases relative to carbon sequestration.

Edge-of-field conservation practices are a much-needed tool for addressing nutrient loading from agricultural systems throughout the Corn Belt Region. Because all the reviewed practices performed equally well with regards to nitrate rention, at least on an efficiency basis, other co-benefits and potential diservices require more attention to inform stakeholder decisions moving forward.

Data availability

Data used for review is included in supplemental table.