|
The Effect of Drainage and Subirrigation From a Small Drainage Water Recycling Reservoir on Corn and Soybean Yields in Eastern North Carolina
Hossam Moursi1, Mohamed A. Youssef1,*, Chad Poole1
Published in Journal of the ASABE 67(1): 13-25 (doi: 10.13031/ja.15536). Copyright 2024 American Society of Agricultural and Biological Engineers.
1Biological and Agricultural Engineering, North Carolina State University, Raleigh, North Carolina, USA.
*Correspondence: mayousse@ncsu.edu
Submitted for review on 15 January 2023 as manuscript number NRES 15536; approved for publication as a Research Article and as part of the “Advances in Drainage: Selected Works from the 11th International Drainage Symposium” Collection by Community Editor Dr. Zhiming Qi of the Natural Resources & Environmental Systems Community of ASABE on 11 October 2023.
Highlights
Abstract. Drainage water recycling (DWR) has been proposed as a source of supplemental irrigation to increase crop production resilience to extended and more frequent dry periods during the crop growing season; however, the system’s potential benefits have not been adequately quantified. The main objective of this study was to assess the performance of a DWR system for providing water for supplemental irrigation to corn and soybean at a research site in eastern North Carolina and quantify corn and soybean yield responses during 4 growing seasons (2018-2021) with varying weather conditions. Two treatments were implemented at the study site: DWR and the control (CT) treatment. The CT treatment was a 11.23 ha non-irrigated field that was primarily drained by a surface drainage system. The DWR treatment (11.48 ha) had a subsurface drainage system that provided drainage during the wet periods and subirrigation during the dry periods of the growing season. A small size reservoir (5,458 m3) was used to collect surface runoff and subsurface drainage and subirrigate the DWR treatment. Results showed that the DWR reservoir stored enough water to meet irrigation requirements in 3 of the four growing seasons and provided 5 to 73 mm of irrigation to the DWR treatment. Subirrigation raised the groundwater table by an average of 15 cm, which helped increase the upward movement of soil water to the root zone and meet crop evapotranspiration demand. DWR increased corn yields by 0.13 and 0.91 Mg ha-1 (1% and 79%) and soybean yields by 0.31 and 0.59 Mg ha-1 (9% and 30%). Subirrigation, which is generally less efficient than overhead irrigation methods, did not optimize the use of the limited water stored in the small reservoir and could not provide enough protection to corn against prolonged dry conditions in the 2019 growing season. The amount of nutrients recycled back to the field through subirrigation was not large enough to help reduce fertilizer application rate. Overall, the results demonstrated that DWR is a promising practice for increasing the resilience of crop production in the southeastern U.S. to the uncertainty in precipitation, which is expected to intensify by climate change. Monitoring the performance of DWR for longer periods with varying factors of weather, soil, and system design and management would help guide the design and management of the system to optimize the performance and minimize the implementation cost.
The world’s population is predicted to increase from 7.9 billion in 2021 to 9.7 billion by 2050 (UN-DESA, 2019). It is estimated that crop production needs to be doubled by 2050 to meet this growing global demand (FAO, 2018; Jaggard et al., 2010; Tilman et al., 2011). However, climate uncertainty and extreme weather conditions present a major challenge for agriculture (Gregory and George, 2011; Lobell et al., 2009). In the United States, climate change projections suggest that water resources and crop production will be significantly affected by rising temperature and highly variable precipitation patterns (Walthall et al., 2013). The increased frequency and duration of dry and wet periods during crop growing seasons intensify plant stress from both excess and deficit soil water conditions, leading to significant crop yield losses.
In humid regions, agricultural drainage is used to improve crop production on naturally poorly drained soils with frequently shallow water table. It reduces plant stress due to excess water conditions during wet periods of the growing season by lowering the water table and improving soil aeration in the plant root zone (Skaggs et al., 1994). Over 40 million ha of cropland in the United States are artificially drained, of which 22 million ha are tile drained and 18 million ha are drained by ditches (USDA-NASS, 2019). The state of North Carolina has approximately 2 million ha of cropland, 37% of which is artificially drained, mostly by field ditches (USDA-NASS, 2019). Improved drainage, however, increases nutrients losses from agricultural fields to downstream surface water (Rabalais et al., 2007; Smith et al., 2015; Helmers et al., 2022).
Corn and soybean are the main crops grown in the US Midwest and the southeastern state of North Carolina. The most critical growth stages of corn are the late vegetative and reproductive stages, while pod development and filling are the most critical stages for soybean (USDA-NRCS, 2010). Soybean can tolerate drought stress until bloom if there is adequate moisture during germination and early seedling growth (USDA-NRCS, 2010). In contrast, corn is sensitive to water deficit conditions, and its yield can significantly drop when water shortage occurs for even a few days, especially during the critical growth stages (USDA-NRCS, 2010; Steduto et al., 2012).
Supplemental irrigation can protect crops in humid regions from severe drought conditions during prolonged periods with precipitation deficit (Troy et al., 2015). Several studies demonstrated that supplemental irrigation in the Southeastern United States increased corn yield by 30%-142%, or 2.30-6.30 Mg ha-1 (Adamsen, 1992; Jordan et al., 2014; Wagger and Cassel, 1993), and soybean yield by 24%-135%, or 0.60-1.10 Mg ha-1 (Ashley and Ethridge, 1978; Doss et al., 1974; Garcia et al., 2010). However, the expanded use of irrigation poses a stress on fresh water sources. In the United States, increased demand for groundwater has led to lowering water table levels and depletion of groundwater aquifers, such as the High Plains aquifer where water levels have declined by more than 30 m since 1950 (Konikow, 2013; McGuire, 2017).
Drainage water recycling (DWR) has the potential to increase crop production, improve water quality, and conserve water resources (Frankenberger et al., 2017; Hay et al., 2021; Moursi et al., 2023). The practice, which may also be referred to as on-farm water storage and reuse, involves capturing and storing drainage water in on-farm reservoirs during wet periods and using this water as a source for supplemental irrigation during the dry periods of the growing season. DWR systems increased corn yield by 1% to 26% and soybean yield by 4% to 17% compared to non-irrigated crops at sites in the U.S. and Canada; however, data were limited to a few sites and influenced by each site-specific conditions, with some sites having only one year of data (Allred et al., 2014; Nelson et al., 2017; Nelson, 2017; Niaghi et al., 2019; Tan et al., 2007). Synthesized data from the Midwestern U.S. showed that DWR increased corn resiliency to precipitation extremes during the critical growth stages of V9-R2 (Willison et al., 2021). The average corn yield increase by DWR was 3.6 Mg ha-1 when precipitation was low (27-85 mm) during V9-R2, compared to non-irrigated corn with only free drainage.
Additional research is needed to expand the knowledge on DWR impacts on crop yield since the published data were based on a limited number of studies with unique site conditions (Hay et al., 2021). The main goal of this study was to experimentally investigate the effect of DWR on corn and soybean yields for a research site in eastern North Carolina, which represents common conditions of the agricultural landscape in the Atlantic Coastal region of the US Southeast. The specific objectives were to: (1) evaluate the effect of subirrigation from a DWR reservoir on corn and soybean yields for 4 growing seasons (2018-2021), (2) investigate the factors affecting crop yield response to DWR, such as weather condition and timing and amount of irrigation, and (3) quantify the amount of nutrients that can be recycled back to the field through irrigation and evaluate its effect on potentially reducing fertilizer application rate.
Materials and Methods
Site Description and Soil Properties
The experimental study was conducted from May 2018 to October 2021 at a privately-owned farm in the Atlantic Coastal Plain region of North Carolina, U.S. (35°29'45.32" N, 76°42'56.85" W; fig. 1). Two experimental treatments were implemented at the study site: drainage water recycling treatment (DWR) and control treatment (CT). The CT treatment (11.23 ha) was a non-irrigated field that was primarily drained by a surface drainage system utilizing shallow open ditches (depth = 50 cm, spacing = 60 m) to collect surface runoff from the treatment. Like the CT treatment, the DWR treatment (11.48 ha) had shallow open ditches for drainage purposes. A subsurface drainage system was designed using the DRAINMOD model (Skaggs et al., 2012) and installed primarily for subirrigating the field. According to DRAINMOD simulations, laterals’ design depth was 125 cm and design spacing was 18 m. Lateral subsurface drains (10 cm in diameter) were laid at a slope of 0.1% while the submains (20 cm in diameter) and the main subsurface drain (25 cm in diameter) were laid at slopes ranging from 0.1% to 0.3% (fig. 1).
Surface runoff from both treatments and from an upstream forested land (32.5 ha) and subsurface drainage from the DWR treatment were stored in a small and elongated reservoir, located between the two treatments. The reservoir had a surface area of 0.38 ha (3840 m2) and a maximum depth of 2.51 m, which provided a storage capacity of 5,458 m3. Only a portion of the surface runoff from both treatments (5.82 ha for the DWR treatment and 4.43 ha for the CT treatment) was routed to the reservoir due to the surface slope. The stored water in the reservoir was used to subirrigate the crops grown on the DWR treatment during the dry periods of the growing season.
The field site has a nearly flat topography (0.11% average slope). According to the USDA Web Soil Survey (USDA-NRCS, 2021), the dominant soil series on the study site is classified as Altavista fine sandy loam, which is a moderately well drained soil. The lateral saturated hydraulic conductivity, determined using the auger hole test, was estimated to be 1.0 cm hr-1. Undisturbed soil core samples, collected from the field and used to develop soil water characteristic curves (Klute and Page, 1986), showed that the soils on the CT and DWR treatments have similar soil properties (fig. S1, table S1). Field investigations indicated the existence of a restrictive soil layer at an approximate depth of 370 cm.
 |
| Figure 1. Location and layout of the study field site. |
Agronomic Practices
Cropping system and field management (e.g., crop rotation, planting and harvesting, tillage practices, fertilizer application, and pest management) were similar for the DWR and CT treatments (table 1). A corn-soybean rotation was implemented at the study site, with soybeans planted in 2018 and 2020, and corn planted in 2019 and 2021. Planting of soybean in 2018 was delayed since it was grown after winter wheat. The wet conditions in the field during the second half of June further delayed soybeans planting until the beginning of July, a month later than the planting date of the 2020 soybeans.
| Table 1. Crop management information for the experimental study site. | |||||||
| Year-Crop | Variety | Population (No. ha-1) | Planting Date | Harvest Date | Fertilizer (N, P, K) | Tillage | |
| Date | Rate (kg ha-1) | ||||||
| 2018-Soybean | AgVenture 56W6R | 321,237 | 8-Jul | 4-Nov | NA | NA | Vertical |
| 2019-Corn | Pioneer 1197 | 74,132 | 26-Apr | 3-Sep | 26-Apr | 32, 23, 0 | Vertical |
| 26-Apr | 18, 26, 0 | ||||||
| 27-May | 179, 0, 0 | ||||||
| 2020-Soybean | MorSoy 5110 | 321,237 | 6-Jun | 5-Nov | 3-Jun | 20, 15, 55 | Vertical |
| 2021-Corn | Pioneer 1197 | 74,132 | 26-Apr | 7-Sep | 19-Apr | 25, 28, 69 | Strip |
| 19-Apr | 20, 29, 0 | ||||||
| 26-May | 137, 0, 0 | ||||||
Drainage and Irrigation Water Management
A water level control structure (with two adjustable weir plates) was installed at the outlet of the main drain of the DWR treatment to regulate water table level and drainage outflow (fig. 2). The weir plates of the control structure were adjusted by a smart system (Bagheri, 2022; Moursi, 2022), which fully automated the change in the structure’s settings based on the field water table depth (WTD). Drainage and subirrigation were triggered automatically based on predefined WTD thresholds that were set depending on crop type and growth stage (table 2). The design of the double-gate assembly of the control structure enabled four modes of operation in response to wet and dry conditions at the field (table 2). The operation modes of the control structure were: (1) free drainage (FD, bottom gate was open) when the field was very wet (WTD < 5 cm); (2) lower controlled drainage (LCD, top gate at 60 cm below soil surface) when WTD was shallower than the optimum depth shown in table 2, but the field was not severely wet; (3) upper controlled drainage (UCD, top gate at 20 cm below soil surface) when WTD was deeper than the optimum depth; and (4) subirrigation (SI, irrigation was applied with the same settings as the UCD) when WTD was deeper than the optimum depth and irrigation was needed. Irrigation was applied by pumping water from the reservoir into a storage tank (11.4 m3) to create water head for subirrigating the field by gravity (fig. 3).
 |
| Figure 2. [a] Water level control structure installed at the outlet of the main drain of the DWR treatment (left of a). A smart drainage system is connected to the control structure to manage drainage and subirrigation (right of a). [b] A picture showing the design of the weir plates that were installed inside the control structure. |
| Table 2. Water table depth thresholds used by the smart system to trigger free drainage (FD), lower controlled drainage (LCD), upper controlled drainage (UCD), and subirrigation (SI) modes.[a] | |||||
| Year (Crop) | DAP[b] | Water Table Depth (WTD; cm) | |||
| FD | LCD | SI | UCD (optimum) | ||
| 2019 (corn) | 0- harvest | 5 | 30 | 60 | 30-60 |
| 2020 (soybean) | 0-34 | 5 | 30 | No irrigation | >30 |
| 34-120 | 5 | 40 | 65 | 40-65 | |
| 120-harvest | 5 | 40 | No irrigation | >40 | |
| 2021 (corn) | 0-40 | 5 | 30 | No irrigation | >30 |
| 40-harvest | 5 | 30 | 65 | 30-65 | |
[a] The system sets the control structure on the upper controlled drainage (UCD) mode if drainage or irrigation was not needed. [b] DAP: Days after planting. Information of the growing season of 2018 is not shown in the table since the smart system was installed in 2019. | |||||
 |
| Figure 3. Water was pumped from the reservoir to a storage tank before it was used for subirrigating the DWR treatment through the subsurface drainage system. |
Data Collection
V-notch weirs were used to measure surface runoff and subsurface drainage flow from the CT and DWR treatments. Water stages upstream and downstream of each weir were measured using Campbell Scientific self-compensating pressure transducers (CS451). The flow rate was calculated using the V-notch weir equation, which estimates flow as a function of water stage and weir dimensions. Liu (2017) provided a detailed description of flow rate calculations. Subirrigation rates were measured using a flowmeter equipped with pulse output (PRM Woltmann Helix style). The flowmeter sent a pulse to a data logger every 0.0038 m3 of irrigation. Subirrigation data recorded by the data logger were calibrated with manual readings of the flowmeter. Flow rates were estimated and stored by CR-200X data loggers (Campbell Scientific) that were installed at each flow measuring station. The time interval of surface runoff and irrigation measurements was 30 min, while all other measurements were hourly.
Nutrient losses from the two treatments and recycled nutrients to the DWR treatment through subirrigation were monitored during the period May 2019-August 2021. The flow measurement stations were instrumented with automated water quality samplers (WS700, Global Water) to collect flow proportional discrete samples for water quality analysis. A detailed description of the water quality monitoring at the study site was provided by Moursi (2022) and Moursi et al. (2023).
Precipitation and air temperature were measured using automated tipping bucket rain gauges (HOBO; Onset). The 30-year average precipitation data (1975-2004) were retrieved from a nearby weather station (Belhaven 3 NE; 35.573° N, -76.5848° W, 14.6 km from the field site). The reservoir water level was measured using an automatic water level data logger (U20-001-01, Onset) installed in a 50-mm diameter PVC observation well. Readings of a manual staff gauge were used to calibrate the water level logger. The WTD in both treatments were measured using automatic water level data loggers (U20-001-01, Onset) installed in 50-mm diameter PVC observation wells midway between the subsurface drains for the DWR treatment and between the open ditches for the CT treatment. Water table data measured by the data loggers were calibrated using manual measurements.
Crop yield data from both treatments were collected at the end of each growing season using a combine harvester with a calibrated yield monitoring system. The combine was equipped with a Global Positioning System (GPS) to record the geographic coordinates of yield data points. Dry yield was calculated by adjusting the yield data to a standard moisture content (15.5% for corn and 13% for soybean). ArcMap software was used to process the yield data, generate the spatial distribution maps, and calculate the average crop yield. The GPS yield data obtained from the harvester underwent cleaning procedures in ArcMap to eliminate the effects of combine turns and overlaps.
Crop Evapotranspiration Under Standard Conditions (ETc)
Daily crop evapotranspiration under standard conditions, or non-stressed, soil water conditions (ETc) was estimated by multiplying grass reference evapotranspiration (ETo) by crop coefficient (Kc). Daily ETo was estimated using the FAO-56 Penman-Monteith method (Allen et al., 1998). Air temperature data were measured at the study site, while wind speed, relative humidity, and solar radiation data were obtained from a nearby weather station (AURO Pamlico Aquaculture Field; 13 km from the site). The Kc for the corn and soybean grown at the study site (fig. S2) were based on values obtained from the North Carolina irrigation guide (USDA-NRCS, 2010).
Irrigation Demand and Reservoir Water Availability
The applied irrigation amounts were compared to irrigation demand and reservoir water availability on a monthly basis to investigate the DWR system’s performance on meeting the irrigation demand. This comparison was conducted for all growing seasons, except for 2018, when reservoir water level data were not available. The monthly irrigation demand was calculated as the difference between precipitation and ETc. If precipitation is greater than ETc, no demand for irrigation was assumed for the month. It should be mentioned that this assumption may not always be valid because the temporal variation of precipitation within the month is another important factor in addition to the total amount of monthly precipitation. In addition, the actual irrigation demand may be higher since only a portion of received precipitation would be stored in the soil profile and be available for crop consumption. Nonetheless, conducting this analysis on a monthly time scale would be useful for demonstrating the interaction and dynamic nature of the supply and demand for water during the growing season as affected by inter- and intra-seasonal variation of precipitation. For the purpose of this analysis, the monthly average of reservoir available water was calculated as a depth using the monthly average of the reservoir water volume divided by the area of the DWR treatment.
Corn and Soybean Growth Stages
The timing and length of growth stages of soybean planted at the study site were obtained from Evans and Fausey (1999). Corn growth stages were determined based on the accumulated growing degree days (GDD) approach (Abendroth et al., 2011). The GDD values were estimated on a daily basis using equation 1.
(1)
where Tmax and Tmin are maximum and minimum daily air temperature, respectively. Tbase is base temperature, below which crop growth stops, which is 10° C for corn and soybean. The growth rate of corn or soybean would increase if temperature is in the range of 10° C to 30° C. Therefore, if Tmax was greater than 30° C, Tmax was set at 30° C, and if Tmin was less than 10° C, Tmin was set at 10° C.
Corn is more susceptible to wet conditions during the establishment through late vegetative stages (V4-V18), and more sensitive to dry conditions during the late vegetative through early reproduction (flowering; V16-R2) period (Evans et al., 1991; Evans and Fausey, 1999). Soybean is more sensitive to wet conditions during the establishment through pod filling stages (V1-R5), while it is more sensitive to dry conditions during the pod developing and filling (R3-R5) stages (Evans et al., 1991; Evans and Fausey, 1999).
Statistical Analysis
Spatial statistical analysis was performed to assess the significance of the differences in corn and soybean yields within and between the CT and DWR treatments. Analysis of variance (ANOVA) was conducted for each crop (corn and soybean), considering two factors and their interactions: (1) the treatment factor with two levels: CT and DWR; and (2) the shallow drainage ditches factor with two levels: near the ditches and midway between ditches. Each treatment was partitioned into 30-m wide transects such that each of the 60-m spaced ditches is located at the center of one of the transects. Therefore, the area between two parallel ditches is divided into one 30-m wide transect located midway between the ditches and two half-transects (15-m wide, each) adjacent to each of the two parallel ditches. By employing ArcGIS, zonal statistics for each transect were calculated from the crop yield raster data. The assumptions of ANOVA, namely, the assumptions of independence, normality, and homogeneity of variances, were thoroughly examined by testing the soybean and corn yield datasets. The analysis revealed that the assumptions were met for the soybean dataset. However, in the case of the corn dataset, it was found that the normality and homogeneity of variances assumptions were not satisfied. To address this, a reciprocal transformation was applied to the corn data. The tests were conducted at a 5% level of significance (a?=?0.05), and all statistical procedures were conducted using RStudio statistical software (R Core Team, 2021; RStudio Team, 2022).
Results and Discussion
Crop Yield Response to Drainage Water Recycling
In 2017, prior to the experimental study, corn yield was similar for both portions of the field where the DWR and CT treatments were implemented (9.16 Mg ha-1 for the CT and 9.11 Mg ha-1 for DWR). Implementing DWR increased corn and soybean yields in all years of the study, compared to the CT treatment (table 3). Long-term averages of corn and soybean yields (1990-2020) planted in Beaufort County, North Carolina, are 7.3 and 2.3 Mg ha-1, respectively (USDA-NASS, 2021). Corn and soybean yields at the DWR treatment exceeded the long-term county average, with the exception of corn in 2019. Average soybean yield at the DWR treatment was 0.45 Mg ha-1 (17%) higher than the yield of non-irrigated soybean, with greater yield increase in 2018 compared to 2020 (table 3). Corn yields of both treatments in 2019 were extremely low due to the severely hot and dry weather conditions. The DWR system increased corn yield in 2019 by 0.91 Mg ha-1, which represents a 79% yield increase compared to the corn grown on the CT treatment. In contrast, corn yields of both treatments in 2021 were high, with DWR increasing corn yield by 0.13 Mg ha-1, which represents only 1% increase compared to the CT treatment.
The spatial yield data collected using the GPS-enabled combine were used to generate spatial yield maps (e.g., fig.4) and perform a spatial statistical analysis. The spatial statistical analysis performed on the yield data showed that the increase in soybean yield at the DWR treatment was statistically significant (table 4; P value = 0.018). In contrast, the increase in corn yield achieved by DWR was not significant (P value = 0.085). The lack of significance of corn yield response to DWR was due in part to the large difference between corn yields in 2019 and 2021, which could have masked the treatment effect. Furthermore, the DWR impact on corn yield was limited due to the lack of water supply in the extremely dry growing season of 2019 and the low demand for irrigation in the relatively wet growing season of 2021 (table 3).
The statistical analysis also showed that the drainage ditches had a significant impact on the spatial yield variability of soybean within each treatment (P value = 0.035). The 30-m wide transects that are adjacent to the ditches had significantly lower soybean yield, compared to the transects that are midway between the ditches (table 4). The transects adjacent to the ditches also had lower corn yield than the transects midway between the ditches (table 4) but the yield difference was not statistically significant (P value = 0.098). The consistent lower crop yield in the areas adjacent to the drainage ditches compared to the areas farther away from the ditches clearly indicates that deficit water conditions were the main cause of stress on the crop compared to excess water conditions. As a result, the observed yield increase in the DWR treatment can mainly be attributed to supplemental irrigation. Subsequently, the difference in the drainage system between the DWR and CT treatments is expected to have a marginal effect on the yield difference between the two treatments. For both crops, the interaction between the treatment and ditch factors was not statistically significant (P value > 0.681), which means the effect of each factor on the crop yield did not depend on the other factor.
| Table 3. Measured corn and soybean yields at the DWR and CT treatments during the 2018-2021 study in eastern North Carolina. | |||||||
| Year | Crop Yield (Mg ha -1) | Yield Increase | |||||
| CT | DWR | Mg ha-1 | % | ||||
| Soybean | 2018 | 1.98 | 2.57 | 0.59 | 30 | ||
| 2020 | 3.36 | 3.67 | 0.31 | 9 | |||
| Mean | 2.67 | 3.12 | 0.45 | 17 | |||
| Corn | 2019 | 1.16 | 2.07 | 0.91 | 79 | ||
| 2021 | 12.27 | 12.41 | 0.13 | 1 | |||
| Mean | 6.72 | 7.24 | 0.52 | 8 | |||
Impact of Weather Conditions on Crop Yield
Annual precipitation during the 4-year study period ranged from 1046 mm in 2019 to 1583 mm in 2018, which represents 10% below and 36% above the 30-year average of 1163 mm, respectively (table 5). The annual precipitation was below the 30-year average in two years (2019 and 2021) and above the 30-year average in two years (2018 and 2020). The 30-year average precipitation during May-October (when corn and soybean are usually grown in North Carolina) was 647 mm, which accounted for 56% of the annual precipitation. During the 4-year study period, May-October precipitation ranged from 554 mm in 2019 (14% < the long-term average) to 873 mm in 2018 (35% > the long-term average). The total precipitation was greater than the total crop evapotranspiration (ETc) during all growing seasons except for 2019, when ETc (544 mm) was 79% greater than the growing season precipitation (303 mm; fig. 5). Although total precipitation during the growing season may be high, it may not occur when crops need it. For example, May-October precipitation was the largest in 2018 compared to the other years. However, monthly ETc during this period was greater than monthly precipitation, except for July (fig. 5). An effective DWR system can store surface and subsurface drainage water during precipitation excess periods to be used later during periods with high evapotranspiration demand. Precipitation during the non-growing season period (November-April), which would determine the reservoir water availability at the beginning of the growing season, ranged from 466 mm in 2021 to 752 mm in 2020.
Moreover, the precipitation from May to July in 2019 was less than the precipitation during the same period across all 30 years. Precipitation during the critical late vegetative period (V9–VT) of corn in 2019 was only 38 mm, while it was 123 mm in 2021 (table 6). Similarly, precipitation during the critical flowering stage through the development stage (V9-R3) of soybean in 2018 was only 16 mm, while it was 165 mm in 2020. Despite large precipitation during the critical pod filling period (R5) of soybean in 2018 (107 mm), almost half of this precipitation (57 mm) occurred on one day during Hurricane Florence. The temporal variation of growing season precipitation can explain the higher yields for both treatments of corn in 2021 compared to 2019, and soybean in 2020 compared to 2018. It also shows that greater yield benefits can be achieved by DWR with lower precipitation during the critical stages, as the DWR yield increase was higher for soybean in 2018 and corn in 2019 (table 3). Willison et al. (2021) also reported that DWR achieved higher corn yield benefits when precipitation was low during the critical period V9-R2.
 |
| Figure 4. An example of the spatial distribution yield maps created for soybean planted in 2020 at the study site. |
The 2019 growing season was not only severely dry, but also substantially hot. Some studies reported that corn yield can be reduced by as high as 94 kg ha-1 for each day the temperature reaches 35°C or higher during the tasseling-silking and grain fill growth stages (Neild and Newman, 1990). In May-July of 2019, Tmax was higher than 35°C in 55 days, and higher than 40°C in 11 days. This combination of hot and dry weather severely affected corn yield of this year.
 |
Irrigation and Reservoir Water Availability
The reservoir water availability at the beginning of the growing season depends on the reservoir storage capacity and the contributing area that provides inflows to the reservoir. The reservoir was full at the beginning of each growing season, except for 2019, when it was partially drained in April to install the control structure at the reservoir outlet (fig. 6). The water stored in the reservoir at the beginning of the 2019 growing season was 3,044 m3, which represents 56% of storage capacity. The reservoir was full at the beginning of the 2021 growing season, which was preceded by the lowest non-growing season (November-April) precipitation compared to the other 3 growing seasons (table 5). The data showed that the site received less precipitation prior to the 2021 growing season than precipitation of the same period (November-April) in 18 of the 30-year historical record. This suggests that the reservoir at this site would be full at the beginning of the growing season in most of the years.
| Table 6. Precipitation during the main crop growth stages of corn and soybean planted at the study site. | |||||
| Year | Precipitation (mm) | ||||
| V1-V6 | V9-R1 | R3 | R5 | R6-R8 | |
| 2018 (Soybean) | 245 | 11 | 5 | 107 | 70 |
| 2020 (Soybean) | 238 | 150 | 15 | 58 | 115 |
| V1-V8 | V9-VT | R1-R2 | R3-R5 | R6 | |
| 2019 (Corn) | 81 | 38 | 51 | 104 | 30 |
| 2021 (Corn) | 149 | 123 | 42 | 198 | 58 |
 |
| Figure 5. Monthly precipitation and crop evapotranspiration (ETc) during the 2018-2021 growing seasons at the study site. |
 |
| Figure 6. Cumulative irrigation and precipitation along with reservoir water level during the growing seasons of 2019, 2020, and 2021 at the study site. The growing season of 2018 is not shown since recording reservoir water level data had not started until the end of the 2018 growing season. |
The amount and timing of irrigation applied to the DWR treatment varied each year, depending on the crop water requirements and water availability in the reservoir (table 7 and fig. 6). The annual amount of irrigation ranged from 5 mm in 2018 to 73 mm in 2020. In 2019, irrigation stopped multiple times after the water level in the reservoir reached the minimum allowable limit (40 cm). Due to the combined dry and hot weather conditions during this growing season and the low reservoir water volume at the beginning of the growing season (56% of reservoir capacity), water in the reservoir was not enough to meet crop water requirements. In contrast, the reservoir met the demand for irrigation in 2020 and 2021. In 2018, irrigation was stopped on 11 September in preparation for Hurricane Florence, which brought 102 mm of precipitation to the study site during 12-18 September. Due to the storm surge that occurred during Hurricane Florence, water in the reservoir had a high salinity level (sodium and chloride levels were 898 ppm and 1320 ppm, respectively) that prevented irrigation resumption. The reservoir had to be pumped dry and refilled by runoff twice to lower the salinity back to levels appropriate for irrigation.
Irrigation demand was highest in 2019 compared to the other years, with 125 mm of irrigation demand in July alone (fig. 7). Irrigation was applied early in the growing season of 2019 with an amount of 19 mm in May to meet the irrigation demand (16 mm) when there was water in the reservoir. During this growing season, the limiting factor affecting the system performance was the reservoir water availability, as the irrigation rate was near the irrigation demand when the reservoir had available water. There was not enough water in the reservoir to continue irrigation during the late vegetative and pollination stages (late June – mid-July) to meet the high evapotranspiration demand during these periods.
 |
| Figure 7. Total irrigation, total irrigation demand (precipitation – crop evapotranspiration), and average reservoir water volume for each month of the growing seasons of 2019, 2020, and 2021. Negative values of irrigation demand mean precipitation was higher than evapotranspiration. |
| Table 7. The amounts and dates of irrigation applied to the DWR treatment at the study site. | |||||
| Crop | Irrigation Amount and Timing[a] | Number of Days When Irrigation Was Applied | |||
| Amount (mm) | Start Date | End Date | |||
| Soybean | 2018 | 5 | 19 July | 11 September | 17 |
| 2020 | 73 | 10 July | 15 September | 53 | |
| Corn | 2019 | 42 | 7 May | 18 August | 40 |
| 2021 | 22 | 30 June | 17 August | 31 | |
[a] Irrigation was applied between the start and end dates, but it was not continuous and depended on crop water requirements and reservoir water availability. | |||||
The irrigation demand in 2020 was low, and irrigation was applied to keep the water table level at a depth that could increase the upward flux to meet crop evapotranspiration. Despite the large monthly precipitation in July, August, and September of 2020, this precipitation occurred in the form of large storm events. For example, 47% of total precipitation in July 2020 occurred as a large storm on only one day (7/15/2020; 59.5 mm). The applied irrigation benefited the crop by alleviating water-deficit stress that would have occurred during the dry periods between these intermittent large storm events. Although the site received a considerable amount of precipitation during the early periods of the growing season (238 mm during 6/6-7/29), precipitation during the critical stage of pod development (R3; 8/19-9/7) was only 15 mm. During this period, 32 mm of irrigation was applied to the DWR treatment to partially meet the high evapotranspiration demand (54 mm). In 2021, the site received a relatively large amount of precipitation in June and July (352 mm). Despite this large precipitation, the demand for irrigation during the critical late vegetative and pollination periods during 6/21-7/23 (V12 – R2) was also high (61 mm).
Two contrasting scenarios can be identified based on these results. The 2020 and 2021 growing seasons represent a scenario with no or small precipitation deficit during the different months of the growing season. In this scenario, the demand for irrigation was low and the relatively small size reservoir adequately met this demand. The 2019 growing season represents the other scenario, which had severe water shortage throughout the growing season. The water stored in the reservoir met the irrigation demand in May before it was used up. The amount of available water in the reservoir was negligible, compared to the large demand for irrigation caused by the severely dry conditions in June and July of 2019. The amount and timing of precipitation during the crop growing season vary from year to year, creating a wide range of distinct scenarios between the two scenarios represented by the 2019, 2020, and 2021 growing seasons.
The performance of a DWR system in modulating crop stresses due to water deficit conditions and subsequently minimizing yield losses depends on the size of the storage reservoir, the length and severity of the dry conditions during the growing season, and the sensitivity of the crop to water shortage, which varies among different crops and among different growth stages for the same crop. Models, such as DRAINMOD-DWR (Moursi et al., 2022), can be extremely useful in understanding the performance of the system as affected by the controlling factors to optimize system performance. The reservoir at the experimental site had a relatively small storage capacity and was not expected to provide enough water to meet the irrigation demand for extended periods of dry conditions during the growing season. For systems with small size reservoirs, irrigation decisions should be carefully made, taking into consideration the current level of available water in the reservoir and the current growth stage of the crop. For example, the 2019 corn could have benefited more from using the limited amount of available water for irrigation during June instead of May, when corn was more sensitive to water deficit conditions. If irrigation water is applied via the subsurface drainage systems (i.e., subirrigation), the depth of the water table should be carefully monitored, and subirrigation should start before the water table becomes too deep to be influenced by the subirrigated water.
 |
| Figure 8. Comparison of daily water table levels in the CT and DWR treatments, irrigation, and precipitation during the growing season of soybean planted in 2018 and 2020. The dashed black line represents the soil surface. |
Field Water Table Depth
Subirrigation works by adding water into the soil profile via the drainage system with the goal of maintaining the water table at a relatively shallow depth, which enhances capillary rise and subsequently helps fulfill the crop water requirements during dry periods of the growing season. Over the course of the four growing seasons of the study, WTD in the DWR treatment was shallower compared to the CT treatment, with an average difference of 15 cm (table S2, figs. 8 and 9). At the start of each growing season, the water table depth was similar in both treatments, ranging from 73 cm in 2020 to 100 cm in 2021. As the growing season progressed, the DWR treatment exhibited a slower rate of water table drawdown compared to the CT treatment in response to subirrigation. For instance, subirrigation applied from May 7 to June 17 of 2019 slowed the water table drawdown in the DWR treatment compared to the CT treatment. During this period, the water table dropped 23 cm (from 89 cm to 112 cm) in the DWR treatment compared to a 47 cm drop in the CT treatment (from 97 cm to 144 cm). By the end of the four growing seasons, the DWR treatment had a 38 cm shallower water table depth compared to the CT treatment (WTD = 93 cm at DWR vs. 131 cm at CT).
Effect of DWR on Recycling Nutrients
Total nitrogen (TN) recycled back to the DWR treatment through irrigation water ranged from 0.8 kg ha-1 to 2.3 kg ha-1, while recycled Total phosphorus (TP) ranged from 0.01 kg ha-1 to 0.3 kg ha-1. Average TN and TP concentrations in the irrigation water were 3.0 and 0.3 mg L-1, respectively. As the results suggest, the amount of nitrogen and phosphorus recycled back to the DWR treatment will not be large enough for producers to reduce fertilizer application rate. Both TN and TP recycled via irrigation were smaller than those drained from the field due to denitrification, dilution, deposition, and other processes occurring in the reservoir. A similar conclusion was reported in a two-year field study by Karki et al. (2018) regarding the insufficiency of recycled nitrate to reduce fertilizer application.
 |
| Figure 9. Comparison of daily water table levels in the CT and DWR treatments, irrigation, and precipitation during the growing season of corn planted in 2019 and 2021. The dashed black line represents the soil surface. |
Conclusion
This experimental study investigated the impact of drainage water recycling (DWR) on corn and soybean yields for four growing seasons (2018-2021) at an agricultural field in eastern North Carolina. The following conclusions can be drawn from the study:
- DWR systems can provide water for supplemental irrigation to increase corn and soybean resilience to uncertainty in growing season precipitation in the U.S. Southeast, including the State of North Carolina. Over the course of the 4-year study, the DWR system increased corn and soybean yields in all growing seasons by an average of 0.52 Mg ha-1 (8%) and 0.45 Mg ha-1 (17%), respectively.
- The performance of DWR in alleviating crop drought stresses and reducing yield losses is directly influenced by availability of water in the storage reservoir to provide irrigation to the crop when needed. The size of the storage reservoir is the key factor that determines water availability for supplemental irrigation. Small size reservoirs with limited water storage capacity cannot meet the crop water demand during extremely dry growing seasons. In this study, the small size reservoir (surface area = 0.38 ha, storage capacity = 5,458 m3) stored enough water to meet irrigation requirements in 3 of the 4 growing seasons. However, the system was less effective in protecting corn against prolonged dry conditions in the 2019 growing season.
- The limited water supply provided by DWR systems requires effective irrigation management to optimize the yield benefits. The method of irrigation, as well as the timing and amount of irrigation, will greatly influence the yield outcome. Irrigation decisions should be carefully made while taking into consideration water availability in the storage reservoir and crop sensitivity to deficit water conditions, which varies among crops and varies with stages of growth/development for the same crop.
- Models, such as DRAINMOD-DWR, are needed to aid the design and management of the DWR systems to maximize performance while minimizing cost.
- The amounts of nitrogen and phosphorus that the DWR system recycle back to the field via irrigation are not large enough to reduce fertilizer application rate.
- Further research is needed to monitor the DWR performance in improving crop production as affected by the DWR reservoir storage capacity, method of irrigation, weather conditions, and different characteristics of the irrigated field (e.g., area, soils, and crops). Furthermore, an economic analysis is needed to optimize DWR performance and minimize implementation and management costs.
Supplemental Material
The supplemental materials mentioned in this article are available for download from the ASABE Figshare repository at: https://doi.org/10.13031/24468367
Acknowledgments
The research presented in this paper was supported by the USDA National Institute of Food and Agriculture under Award No. 2015-68007-23193, “Managing Water for Increased Resiliency of Drained Agricultural Landscapes.
References
Abendroth, L. J., Elmore, R. W., Boyer, M. J., & Marlay, S. K. (2011). Corn growth and development (PMR 1009). Ames, IA: Iowa State University.
Adamsen, F. J. (1992). Irrigation method and water quality effects on corn yield in the mid-atlantic coastal plain. Agron. J., 84(5), 837-843. https://doi.org/10.2134/agronj1992.00021962008400050015x
Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). FAO Irrigation and Drainage Paper No. 56 - Crop Evapotranspiration. Rome, Italy: United Nations FAO.
Allred, B. J., Gamble, D. L., Clevenger, W. B., LaBarge, G. A., Prill, G. L., Czartoski, B. J.,... Brown, L. C. (2014). Crop yield summary for three wetland reservoir subirrigation systems in northwest Ohio. Appl. Eng. Agric., 30(6), 889-903. https://doi.org/10.13031/aea.30.10501
Ashley, D. A., & Ethridge, W. J. (1978). Irrigation effects on vegetative and reproductive development of three soybean cultivars. Agron. J., 70(3), 467-471. https://doi.org/10.2134/agronj1978.00021962007000030026x
Bagheri, S. (2022). Managing water table for increasing resilience of drained agriculture to weather variability: System development, numerical, and data-driven modeling. PhD diss. Raleigh, NC: North Carolina State University, Biological and Agricultural Engineering.
Doss, B. D., Pearson, R. W., & Rogers, H. T. (1974). Effect of soil water stress at various growth stages on soybean yield. Agron. J., 66(2), 297-299. https://doi.org/10.2134/agronj1974.00021962006600020032x
Evans, R. O., & Fausey, N. R. (1999). Effects of inadequate drainage on crop growth and yield. In R. W. Skaggs, & J. van Schilfgaarde (Eds.), Agricultural Drainage (Vol. 38, pp. 13-54). American Society of Agronomy, Crop Science Society of America, Soil Science Society of America. https://doi.org/10.2134/agronmonogr38.c2
Evans, R. O., Skaggs, R. W., & Sneed, R. E. (1991). Stress day index models to predict corn and soybean relative yield under high water table conditions. Trans. ASAE, 34(5), 1997-2005. https://doi.org/10.13031/2013.31829
FAO. (2018). How to feed the world. High Level Expert Forum - How to Feed World. Island Press/Center for Resource Economics. https://doi.org/10.5822/978-1-61091-885-5
Frankenberger, J., Reinhart, B., Nelson, K., Bowling, L., Hay, C., Youssef, M.,... Allred, B. (2017). Questions and answers about drainage water recycling for the Midwest. Purdue Extension ABE-156-W. West Lafayette, IN: Purdue University. Retrieved from https://www.extension.purdue.edu/extmedia/ABE/ABE-156-W.pdf
Garcia y Garcia, A., Persson, T., Guerra, L. C., & Hoogenboom, G. (2010). Response of soybean genotypes to different irrigation regimes in a humid region of the southeastern USA. Agric. Water Manag., 97(7), 981-987. https://doi.org/10.1016/j.agwat.2010.01.030
Gregory, P. J., & George, T. S. (2011). Feeding nine billion: The challenge to sustainable crop production. J. Exp. Bot., 62(15), 5233-5239. https://doi.org/10.1093/jxb/err232
Hay, C. H., Reinhart, B. D., Frankenberger, J. R., Helmers, M. J., Jia, X., Nelson, K. A., & Youssef, M. A. (2021). Frontier: Drainage water recycling in the humid regions of the U.S.: Challenges and opportunities. Trans. ASABE, 64(3), 1095-1102. https://doi.org/10.13031/trans.14207
Helmers, M. J., Abendroth, L., Reinhart, B., Chighladze, G., Pease, L., Bowling, L.,... Strock, J. (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
Jaggard, K. W., Qi, A., & Ober, E. S. (2010). Possible changes to arable crop yields by 2050. Philos. Trans. R. Soc. B: Biol. Sci., 365(1554), 2835-2851. https://doi.org/10.1098/rstb.2010.0153
Jordan, D. L., Johnson, P. D., Grabow, G. L., & Corbett, T. (2014). Corn, cotton, and peanut response to tillage and subsurface drip irrigation in North Carolina. Agron. J., 106(3), 962-967. https://doi.org/10.2134/agronj13.0486
Karki, R., Tagert, M. L., & Paz, J. O. (2018). Evaluating the nutrient reduction and water supply benefits of an on-farm water storage (OFWS) system in East Mississippi. Agric. Ecosyst. Environ., 265, 476-487. https://doi.org/10.1016/j.agee.2018.06.024
Klute, A., & Page, A. L. (1986). Methods of soil analysis. Part 1. Physical and mineralogical methods; Part 2. Chemical and microbiological properties. American Society of Agronomy, Inc. Soil Science Society of America, Inc. https://doi.org/10.2136/sssabookser5.1.2ed
Konikow, L. F. (2013). Groundwater depletion in the United States (1900-2008). Scientific Investigations Report, 2013-5079. 75. U.S. Department of the Interior, U.S. Geological Survey. https://doi.org/10.3133/sir20135079
Liu, Y. (2017). The performance of controlled drainage and inline denitrifying woodchip bioreactor for reducing nutrient losses from subsurface drained grassland receiving liquid swine lagoon effluent. PhD diss. Raleigh, NC: North Carolina State University, Biological and Agricultural Engineering.
Lobell, D. B., Cassman, K. G., & Field, C. B. (2009). Crop yield gaps: Their importance, magnitudes, and causes. Annu. Rev. Environ. Resour., 34(1), 179-204. https://doi.org/10.1146/annurev.environ.041008.093740
McGuire, V. L. (2017). Water-level and recoverable water in storage changes, High Plains aquifer, predevelopment to 2015 and 2013–15. Scientific Investigations Report 2017-5040. 24. U.S. Department of the Interior, U.S. Geological Survey. https://doi.org/10.3133/sir20175040
Moursi, H. (2022). Investigating crop yield and water quality benefits of drainage water recycling. PhD diss. Raleigh, NC: North Carolina State University, Biological and Agricultural Engineering.
Moursi, H., Youssef, M. A., & Chescheir, G. M. (2022). Development and application of DRAINMOD model for simulating crop yield and water conservation benefits of drainage water recycling. Agric. Water Manag., 266, 107592. https://doi.org/10.1016/j.agwat.2022.107592
Moursi, H., Youssef, M. A., Poole, C. A., Castro-Bolinaga, C. F., Chescheir, G. M., & Richardson, R. J. (2023). Drainage water recycling reduced nitrogen, phosphorus, and sediment losses from a drained agricultural field in eastern North Carolina, U.S.A. Agric. Water Manag., 279, 108179. https://doi.org/10.1016/j.agwat.2023.108179
Neild, R. E., & Newman, J. E. (1990). Growing season characteristics and requirements in the corn belt. NCH-40. West Lafayette, IN: Purdue University Cooperative Extension Service.
Nelson, K. A. (2017). Soybean yield variability of drainage and subirrigation systems in a claypan soil. Appl. Eng. Agric., 33(6), 801-809. https://doi.org/10.13031/aea.12276
Nelson, K., Harder, D., & Mungyeko-Mayola, M. (2017). Corn response to drainage water recycling in a blackoar silt loam - Greenley Memorial Research Center field day report. Novelty, MO.
Niaghi, A. R., Jia, X., Steele, D. D., & Scherer, T. F. (2019). Drainage water management effects on energy flux partitioning, evapotranspiration, and crop coefficients of corn. Agric. Water Manag., 225, 105760. https://doi.org/10.1016/j.agwat.2019.105760
R Core Team. (2021). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org/
Rabalais, N. N., Turner, R. E., Sen Gupta, B. K., Boesch, D. F., Chapman, P., & Murrell, M. C. (2007). Hypoxia in the northern Gulf of Mexico: Does the science support the Plan to Reduce, Mitigate, and Control Hypoxia? Estuaries Coasts, 30(5), 753-772. https://doi.org/10.1007/BF02841332
RStudio Team. (2022). RStudio: integrated development environment for R. Boston, MA: RStudio, PBC.
Skaggs, R. W., Brevé, M. A., & Gilliam, J. W. (1994). Hydrologic and water quality impacts of agricultural drainage. Crit. Rev. Environ. Sci. Technol., 24(1), 1-32. https://doi.org/10.1080/10643389409388459
Skaggs, R. W., Youssef, M. A., & Chescheir, G. M. (2012). DRAINMOD: Model use, calibration, and validation. Trans. ASABE, 55(4), 1509-1522. https://doi.org/10.13031/2013.42259
Smith, D. R., King, K. W., & Williams, M. R. (2015). What is causing the harmful algal blooms in Lake Erie? J. Soil Water Conserv., 70(2), 27A-29A. https://doi.org/10.2489/jswc.70.2.27A
Steduto, P., Hsiao, T. C., Fereres, E., & Raes, D. (2012). Crop yield response to water - FAO Irrigation and Drainage Paper No. 66. Rome, Italy: United Nations FAO.
Tan, C. S., Zhang, T. Q., Drury, C. F., Reynolds, W. D., Oloya, T., & Gaynor, J. D. (2007). Water quality and crop production improvement using a wetland-reservoir and draining/subsurface irrigation system. Can. Water Resour. J., 32(2), 129-136. https://doi.org/10.4296/cwrj3202129
Tilman, D., Balzer, C., Hill, J., & Befort, B. L. (2011). Global food demand and the sustainable intensification of agriculture. Proc. Natl. Acad. Sci., 108(50), 20260-20264. https://doi.org/10.1073/pnas.1116437108
Troy, T. J., Kipgen, C., & Pal, I. (2015). The impact of climate extremes and irrigation on US crop yields. Environ. Res. Lett., 10(5), 054013. https://doi.org/10.1088/1748-9326/10/5/054013
UN-DESA. (2019). World population prospects 2019. New York: United Nations - Department of Economic and Social Affairs.
USDA-NASS. (2019). The 2017 Census of Agriculture. United States Summary and Data. AC-17-A-51. USDA, National Agricultural Statistics Service.
USDA-NASS. (2021). USDA National Agricultural Statistics Service (USDA-NASS). Retrieved from https://www.nass.usda.gov
USDA-NRCS. (2010). North Carolina irrigation guide. National Engineering Handbook Part 652. North Carolina: USDA-NRCS. Retrieved from https://efotg.sc.egov.usda.gov/references/public/NC/NC_Irrigation_Guide_Apr_2010.pdf
USDA-NRCS. (2021). Web Soil Survey (WSS). Natural Resources Conservation Service, United States Department of Agriculture. Retrieved from http://websoilsurvey.nrcs.usda.gov/
Wagger, M. G., & Cassel, D. K. (1993). Corn yield and water-use efficiency as affected by tillage and irrigation. Soil Sci. Soc. Am. J., 57(1), 229-234. https://doi.org/10.2136/sssaj1993.03615995005700010040x
Walthall, C. L., Hatfield, J., Backlund, P., Lengnick, L., Marshall, E., Walsh, M.,... Ziska, L. H. (2013). Climate change and agriculture in the united states: Effects and adaptation. USDA Technical Bulletin 1935. U.S. Department of Agriculture, University Corporation for Atmospheric Research, National Center for Atmospheric Research.
Willison, R. S., Nelson, K. A., Abendroth, L. J., Chighladze, G., Hay, C. H., Jia, X.,... Wikle, C. K. (2021). Corn yield response to subsurface drainage water recycling in the midwestern United States. Agron. J., 113(2), 1865-1881. https://doi.org/10.1002/agj2.20579