Article Request Page ASABE Journal Article
Weather, Landscape, and Management Effects on Nitrate and Soluble Phosphorus Concentrations in Subsurface Drainage in the Western Lake Erie Basin
L. A. Pease, N. R. Fausey, J. F. Martin, L. C. Brown
Published in Transactions of the ASABE 61(1): 223-232 (doi: 10.13031/trans.12287). 2018 American Society of Agricultural and Biological Engineers.
Submitted for review in February 2017 as manuscript number NRES 12287; approved for publication as part of the “Advances in Drainage: Selected Works from the 10th International Drainage Symposium” collection by the Natural Resources & Environmental Systems Community of ASABE in October 2017.
Mention of company or trade names is for description only and does not imply endorsement by the USDA. The USDA is an equal opportunity provider and employer.
The authors are Lindsay A. Pease, Research Agricultural Engineer, USDA-ARS Soil Drainage Research Unit, Columbus, Ohio (formerly Graduate Student, Department of Food, Agricultural, and Biological Engineering, The Ohio State University, Columbus, Ohio); Norman R. Fausey, Soil Scientist, USDA-ARS Soil Drainage Research Unit, Columbus, Ohio; Jay F. Martin, Professor, Department of Food, Agricultural, and Biological Engineering and Ohio Sea Grant, and Larry C. Brown, Professor, Department of Food, Agricultural, and Biological Engineering, The Ohio State University, Columbus, Ohio. Corresponding author: Lindsay A. Pease, USDA-ARS, 590 Woody Hayes Dr., Columbus, OH 43210; phone: 419-290-6150; e-mail: firstname.lastname@example.org.
Abstract. Subsurface drainage, while an important and necessary agricultural production practice in the Midwest, contributes nitrate (NO3-N) and soluble phosphorus (P) to surface waters. Eutrophication (i.e., excessive enrichment of surface water by NO3-N and soluble P) supports harmful algal blooms in receiving waters. The magnitude of NO3-N and soluble P loss in subsurface drainage varies greatly by landscape, weather, and field management factors. This study evaluated both the relative and combined impacts of these factors on observed NO3-N and soluble P concentrations in subsurface drainage water in the Western Lake Erie Basin watershed. Water quality data from multiple drainage outlet sites in northwest Ohio provided evidence that the primary management factors affecting NO3-N and soluble P loss were the amount and time of fertilizer application. Results strongly support following Tri-State fertilizer recommendations and 4R nutrient stewardship principles to reduce the risk of NO3-N and soluble P loss. Results also provided evidence of NO3-N and soluble P transport to subsurface drains via different pathways. Due to differences in NO3-N and soluble P transport through the soil profile (via baseflow and preferential flow, respectively), management approaches taken to reduce one nutrient may exacerbate losses of the other. Further research is needed to address potential changes in field hydrology (and consequently the in-field transport of soluble nutrients) from different types of agricultural best management practices (BMPs) and to evaluate optimal stacking of BMPs to achieve reductions in both NO3-N and soluble P loss. Controlled drainage has a high potential for stacking with other BMPs because it is primarily a physical discharge and load reduction practice.
Keywords.Agriculture, Eutrophication, Nutrient transport, Regression analysis, Water quality.
Worldwide, more than 415 coastal areas are impacted by eutrophication (Selman et al., 2008) (i.e., the process of becoming more nutrient-enriched). Eutrophication is a serious environmental concern that threatens both aquatic ecosystems and public health. It can lead to limited use of water resources for recreation, fishing, and drinking. Eutrophication occurs when high levels of nutrients, particularly nitrogen and phosphorus, are present in water bodies. Nitrogen and phosphorous are essential nutrients for crop growth, but even relatively small losses of these nutrients from agricultural lands can lead to excessive algal growth in the water bodies of agricultural watersheds.
The emergence of algal blooms in the Gulf of Mexico has been linked to agricultural intensification and increases in fertilizer use (Turner and Rabalais, 2003). Because nitrogen (N) is the primary limiting nutrient for algal growth in seawater, algal blooms in the Gulf of Mexico and Chesapeake Bay have led to a strong focus on reducing agricultural losses of nitrate (NO3-N) across the Corn Belt. At the same time, the growing number of algal blooms in freshwater systems, such as Lake Erie, has called attention to the need to control phosphorus from agricultural runoff. Phosphorus (P) is the primary limiting nutrient for algal growth in freshwater systems. Recent algal blooms in Lake Erie have been correlated with high loads of soluble P from agricultural nonpoint sources (OLEPTF, 2010; Rucinski et al., 2010; Daloglu et al., 2012).
In the Western Lake Erie Basin watershed, 72% of land use is row-crop agriculture, and over 60% of cropland contains subsurface drainage systems (OLEPTF, 2010). To promote settlement and economic development in this and similar adjacent areas, federal funds were appropriated in the mid-1800s to facilitate construction of drains (Beauchamp, 1987). Subsurface drainage, which began in the Midwest in the early 1900s, remains a necessary and common practice across the Western Lake Erie Basin. Closer lateral spacing and shallower drain depths improve crop yields in this region, providing an economic incentive for farmers to continue intensifying drain installations over time. The Western Lake Erie Basin remains one of the most intensively drained regions in the U.S. (Jaynes and James, 2007; OLEPTF, 2010).
Subsurface drainage allows high-productivity agriculture in much of this region, but it also provides a pathway for short-circuiting of water and dissolved nutrients through the soil profile to drainage ditches and streams. Williams et al. (2015a) found that subsurface drains exported 62% of NO3-N from a headwater watershed in Ohio. Within the Maumee River basin, subsurface drains exported 49% of P from four monitored fields in the St. Joseph River watershed (Smith et al., 2015). Increased drainage intensity has been reported to increase the nitrate loads and discharge volume from subsurface drains (Kladivko et al., 2004; Sands et al., 2008). As drainage intensity increases across the Western Lake Erie Basin watershed, nutrient losses from subsurface drains will likely also increase, posing a serious risk to Lake Erie in the future. Understanding the nutrient loss patterns in subsurface drainage could help to mitigate future losses by (1) providing insight about when elevated concentrations of P and NO3-N are expected in subsurface drainage and (2) determining which BMPs are most beneficial for improving drainage water quality.
Soluble P and NO3-N concentrations in subsurface drainage are influenced by a wide range of factors, including soil characteristics, local weather conditions, field hydrology, drain design, field management practices, and drainage management practices (Carpenter et al., 1998; Moog and Whiting, 2002; King et al., 2015b). These factors have often been studied one or two at a time, leaving their relative and combined impacts on dissolved nutrient concentrations unexplored.
The objective of this study was to evaluate the contributions of site, management, and weather variables to the NO3-N and soluble P concentrations found in subsurface drainage water. This study used a multiple linear regression approach to evaluate both the relative and combined impacts of a suite of site-specific factors on soluble P and NO3-N concentrations. Similar approaches were used by Moog and Whiting (2002) and Sprague and Gronberg (2012) to explore variations in riverine discharge of nitrogen and phosphorus based on watershed factors and limited agricultural management details. This work is able to examine variability in nutrient loss on a finer scale than previous studies. By using site-specific details and concentrations, the contributing factors can be assessed for their direct impacts on variations in soluble P and NO3-N concentrations in subsurface drainage. This allows for better identification of risk factors for soluble P and NO3-N losses in the Western Lake Erie Basin.
This study was conducted on eight privately owned field sites in northwest Ohio from 2011 to 2014 (fig. 1). All sites are representative of typical soil properties and drainage system design within the Western Lake Erie Basin. Each site was an agricultural field with two separate subsurface drainage systems under different portions of the field. One outlet at each site was managed with controlled drainage, and the other was conventionally drained (i.e., the outlet was always unmanaged). Management of controlled drainage was guided by the following recommendations but could vary by site based on farmer preferences. In general, adjustable boards in the outlet control structure of the controlled drainage site were set to raise the drainage outlet to a nominal depth of 0.3 m below the field surface when no crop was in the field and to 0.5 m below the field surface during the growing season. The outlet level was set to allow unrestricted drainage before and during planting and harvest periods to ensure adequate drainage to allow farm machinery into the field. Aside from drainage, all other field management operations (planting, harvest, fertilizer application, tillage) were the same for the two subfields at each site. Detailed descriptions of site characteristics, drainage management, and field management are available in Pease (2016).
Figure 1. Approximate site locations (black squares) in the Western Lake Erie Basin watershed (hatched area).
Rainfall was monitored at each site using a tipping-bucket gauge to record cumulative rainfall volume. Missing rain gauge data were estimated on a daily basis using the inverse distance weighting (IDW) method with available rainfall data recorded at locations within a 30 km radius of the field site (Ashraf et al., 1997). Rainfall observations for use with the IDW method were obtained from two sources: other sites included in this study, and available weather station data from the NCDC Global Historical Climatology Network (GHCN) - Daily database, version 3 (Menne et al., 2012). Subsurface drainage samples were taken from each control structure during field visits. Samples were captured and stored in 250 mL prewashed plastic bottles. Sampling typically occurred monthly throughout the year but varied by site depending on the frequency of field visits. Only samples taken during periods of active flow were included in the statistical analysis. If a sample was taken on a day for which submerged outlet conditions were observed or suspected, the sample was excluded from the dataset. A total of 324 soluble P and NO3-N samples from the four-year period were available for statistical analysis. Statistical analysis of weather, landscape, and management factors depended on having all factors represented for each sample. Due to missing information regarding fertilizer application, it was necessary to use a subset of the 324 concentration samples for all factors to be represented for each sample. A subset of 275 concentration samples had adequate N fertilizer application data for inclusion in the NO3-N concentration prediction model. A subset of 237 samples had adequate P fertilizer application data for inclusion in the soluble P prediction model.
Laboratory Analysis of Soluble P and NO3-N Concentrations
Water samples were stored below 4°C and vacuum filtered through a 0.45 µm pore diameter membrane filter prior to analysis. Nitrate (NO3-N) and soluble P (PO4-P) concentrations were analyzed according to U.S. EPA guidelines (USEPA, 1983) and were determined using flow injection analysis with a QuikChem 8000 FIA automated ion analyzer (Lachat Instruments, Milwaukee, Wisc.). NO3-N concentration was measured using copper-cadmium reduction. Soluble P was measured using ascorbic acid reduction (Murphy and Riley, 1962).
Weather, Soil, Drainage System Design, Field Management, and Drainage Management
Twelve main factors were evaluated for their impact on soluble P and NO3-N concentrations. Rainfall, daily temperatures, soil texture, topography, drain spacing, fertilizer rate, and fertilizer timing were represented as continuous factors. Season, soil test P, crop, tillage, and drainage management were represented as discrete factors. Two-way interaction effects were evaluated between all main factors for this analysis.
Three rainfall factors were evaluated for their impact on soluble nutrient concentration. “Daily rainfall” is a measurement of the total amount of rainfall observed on the day of the grab sample. “Previous day rainfall” is a measurement of the total amount observed on the day before the grab sample. “Weekly rainfall” is a measurement of the total cumulative rainfall observed on the six days prior to when the grab sample was taken and on the day the grab sample was taken.
“Minimum temperature” refers to the observed minimum temperature on the day of the sample. These data were not recorded on site but were estimated using data from the nearest weather station, as obtained from the GHCN-Daily database (Menne et al., 2012).
Seasonal variation was captured using “Season” as a quarterly variable. Winter, spring, summer, and autumn were defined as January 1 to March 31, April 1 to June 30, July 1 to September 30, and October 1 to December 31, respectively.
Soil texture was estimated using data from the USDA-NRCS Web Soil Survey. The individual impacts of the percentages of sand and clay on soluble P and NO3-N could not be estimated using multiple linear regression due to the high collinearity of the clay and sand contents at these sites. For NO3-N concentration analysis, “sand content” was used as the predictor representing soil texture. For soluble P analysis, clay and sand content were combined into one index: “clay:sand ratio.” This index was calculated by dividing the percent clay by the percent sand and transformed using natural log.
Topographic variation was estimated using light detection and ranging (LiDAR) digital elevation model (DEM) data ob-tained from the Ohio Statewide Imagery Program. The LiDAR imagery was collected between 2011 and 2013 and was accurate to ±0.76 m. “Site relief” was estimated as the maximum difference in elevation within each subfield.
“Soil test P” (STP) refers to the extractable concentration of P in the soil profile. Values were reported by the farmers as part of a general soil fertility analysis of the field sites. All sites were classified to have “high STP” (21 to 30 mg kg-1) or “very high STP” (>30 mg kg-1) according to fertilizer recommendations for corn in Ohio (Vitosh et al., 1995).
“Drain spacing” refers to the distance between lateral subsurface drains at each field site. Due to a lack of variation in drain spacing across sites (one with 10 m spacing, five with 12 m spacing, and two with 15 m spacing), the sites were divided into two groups: 10 to 12 m spacing and 15 m spacing.
“Drainage water management” (DWM) indicates whether the subfield was under controlled drainage management or conventional drainage management.
“Tillage” refers to the tillage practice employed by the farmer. “Rotational tillage” refers to chisel plow tillage prior to corn planting but no tillage prior to planting soybeans. This approach is the most common tillage practice employed by farmers in the Western Lake Erie Basin. “No-till” refers to continuous no-tillage planting practices. It is estimated that about 20% of farmers in the Western Lake Erie Basin use continuous no-till (USDA-NRCS, 2015).
“Crop” refers to the crop that a farmer grew during the growing season of a given study year. For the purposes of this analysis, the crop grown at a site during a given study year was applied to all samples beginning on May 1 to coincide with typical planting dates and in-field operations and continued through April 30 of the following year to capture any residual impact of the crop during the non-growing season. All farmers in this study followed a corn-soybean rotation with occasional years of winter wheat. Only one farmer grew wheat during the study period (St. Johns site during the 2012 growing season). Two farmers grew popcorn during the 2011 growing season.
Fertilizer timing was represented by “days since N application” for N fertilizer application and “days since P application” for P fertilizer applications. For each sample, this factor equaled the total number of days from the date of N or P fertilizer application to the date of the sample. Fertilizer rate was represented by “N fertilizer rate” for N application and by “P fertilizer rate” for P application. These values represent the rate of elemental N or P applied to the field during the growing season and are cumulative over the growing season to reflect multiple fertilizer applications. Fertilizer application data were obtained from field management notes provided by the farmers. The farmers generally provided application dates and the total amount of N or P applied on each date. In some cases, the farmers’ notes did not provide adequate context to determine the amount or timing of fertilizer application. When the fertilizer records were consistent from year to year, fertilizer application records prior to the start of the four-year study period were used to provide additional context for the fertilizer data. If missing fertilizer application details could not be filled in from prior records, the samples were excluded from further analysis.
The influence of site, weather, and management factors and their interactions on soluble P and NO3-N concentrations were evaluated using a multiple linear regression approach. The concentration values for the response variables “soluble P” and “NO3-N” were log-transformed to be approximately normally distributed prior to analysis. Continuous factors were centered and coded prior to regression analysis to make the variables dimensionless and to improve the interpretability of the statistical tests. On average, five grab samples per drainage treatment at each site were taken on days with discharge each year. With a sampling frequency of fewer than 12 samples per year, it was necessary to adjust for seasonal variation but not for serial correlation in the concentration sample analysis (Loftis and Ward, 1980). Main factors were evaluated for collinearity using variance inflation factor (VIF) testing. Any terms found to be collinear with VIF > 5 were not included in the model.
Statistical analyses were carried out in JMP 11.0 (SAS, 2013). Initial selection of model terms was conducted using a forward selection procedure with the minimum Corrected Akaike’s Information Criterion (AICc) to define the “optimal” model (Akaike, 1974; Burnham and Anderson, 2004) using Stepwise Fit within the Fit Model Platform in JMP. This procedure systematically evaluated factors for inclusion in the model and was used to improve the model’s goodness of fit while adjusting for increased model complexity to reduce the probability of overfitting the model. Further reduction of model terms was required to remove terms that were not statistically significant at a < 0.05. This was conducted manually with backward selection in the Fit Model Platform in JMP.
The observed annual rainfall for northwest Ohio was compared between the study period and the 30-year (1985 to 2014) average. As a baseline, the 30-year annual average rainfall for northwest Ohio was 915 mm. Two years of the study period were observed to have higher than average rainfall: 2011 (1300 mm) and 2013 (970 mm). However, the other two years were observed to have lower than average rainfall: 2012 (780 mm) and 2014 (854 mm) (NCDC, 2016). On a seasonal basis, near-normal amounts of rainfall were observed in all years during the winter. Spring rainfall was greater in 2011 than in any other year from 1985 to 2014. This was followed by higher than average rainfall in the summer and autumn of 2011. The observed rainfall for spring 2012 was approximately half of what was expected from the 30-year average for spring rainfall. Mean daily rainfall across the sites was 2 mm ±6 mm, and mean weekly rainfall was 15 mm ±17 mm (fig. 2).
Observed NO3-N and Soluble P Concentrations
During the four-year study period, the overall mean concentration of NO3-N for both drainage treatments was 8.4 ppm. Seasonally, the mean NO3-N concentrations were 7.9 ppm in winter, 10 ppm in spring, 6.9 ppm in summer, and 5.6 ppm in autumn. During the four-year study period, the overall mean concentration of soluble P was 0.039 ppm. Seasonally, the mean soluble P concentrations were 0.029 ppm in winter, 0.048 ppm in spring, 0.046 ppm in summer, and 0.033 ppm in autumn. The NO3-N and soluble P concentrations over time revealed underlying seasonal trends (fig. 3). The strongest seasonal relationship occurred for both nutrients during summer and autumn. NO3-N decreased throughout the summer and began increasing in the fall. Soluble P increased throughout the summer and decreased in the fall.
Figure 2. Probability of exceedance for daily and weekly rainfall based on rainfall observed across eight field sites in Ohio from 2011 to 2014.
Figure 3. Observed NO3-N and soluble P concentrations in subsurface drainage by day of year during the 2011 to 2014 study period. Data were included from all sites, years, and drainage treatments when drains were actively flowing. Points represent concentrations, and lines represent mean concentrations. The shaded area around each line represents the 95% confidence interval of mean concentration.
Predictive Model Results
Weather, landscape, and management factors all had statistically significant main effects (table 1) and interaction effects (table 2) on NO3-N and soluble P concentrations in subsurface drainage. The final models for estimating NO3-N and soluble P concentrations had R2 values that indicated the models were able to explain 70% and 74%, respectively, of the variability in the dataset. The magnitude of each factor indicates its relative strength in affecting NO3-N or soluble P concentration in comparison to other factors. Values for NO3-N and soluble P concentration were log-transformed prior to analysis, so the change in concentration, as indicated by the magnitude of each factor estimate, is relative rather than absolute. The sign of each factor indicates an increase (+) or decrease (-) in concen-tration compared to the reference case. Effect coding was used for discrete factors, and one factor was coded as a dummy variable as a baseline for statistical comparisons to avoid collinearity. The reference cases were “autumn” for season, “10 or 12 m spacing” for drain spacing, “high” for STP, “corn” for crop, “rotational” for tillage, and “conventional drainage” for DWM.
Table 1. Relative effects of main factors on soluble NO3-N and P concentrations in subsurface drainage.[a] Site Characteristics and Factors NO3-N
Rainfall Weekly rainfall 1.33 NS Previous day rainfall NS 0.72 Daily rainfall 0.87 0.54 Temperature Daily minimum NS NS Season Winter NS NS Spring 0.24 NS Summer NS 0.73 Autumn 0 0 Soil texture Clay:sand ratio - -0.38 Sand content -0.34 - Topography Site relief NS NS Drain spacing 15 m spacing 0.13 NS Soil test P (STP) Very high (STP = 31 ppm) 0.34 0.11 Fertilizer rate and timing N fertilizer rate NS - P fertilizer rate - 1.41 Days since N application -0.59 - Days since P application - NS Crop Popcorn NS NS Soybeans NS 0.66 Wheat -0.05 0.41 Corn 0 0 Tillage No-till -0.17 NS Drainage water management (DWM) Controlled drainage NS NS Intercept 1.35 0.26
[a] NS indicates that the factor was not significant at p < 0.05, a value of 0 indicates that the factor was the dummy variable for effect coding of discrete factors, and a dash (-) indicates that the factor was not included in the model due to collinearity.
Table 2. Interactions of weather, landscape, and management factors on NO3-N and soluble P concentrations in subsurface drainage.[a] Interaction Effects NO3-N
Weather × Weather Interactions Daily rainfall × Weekly rainfall 1.27 NS Daily rainfall × Temperature -1.06 NS Previous day rainfall × Temperature 0.97 NS Winter × Daily rainfall -0.16 NS Previous day rainfall × Summer NS 0.59 Winter × Temperature NS 0.09 Management × Management Interactions Days since N application × Popcorn -0.53 NS No-till × Days since N application -0.29 NS Days since N application × N rate applied -0.23 NS No-till × N rate applied 0.12 NS DWM × N rate applied 0.04 NS Management × Landscape Interactions N rate applied × Sand content 0.72 NS N rate applied × Drain spacing -0.24 NS Days since N application × Drain spacing 0.20 NS N rate applied × Very high STP -0.09 NS No-till × Site relief NS -0.32 No-till × Very high STP 0.07 0.29 Soybean × Site relief 0.06 NS P rate applied × Clay:sand ratio NS 0.09 DWM × Site relief NS -0.07 DWM × Very high STP NS -0.07 Management × Weather Interactions Popcorn × Weekly rainfall 0.26 NS Days since N application × Spring 0.25 NS N rate applied × Daily rainfall -0.21 NS N rate applied × Temperature -0.10 NS No-till × Temperature 0.10 NS P rate applied × Previous day rainfall NS 1.34 Soybean × Previous day rainfall NS 0.46 Wheat × Daily rainfall NS 0.33 Soybean × Daily rainfall NS 0.18 No-till × Weekly rainfall NS -0.11 Days since P application × Winter NS 0.11 Wheat × Summer NS 0.10 No-till × Winter 0.10 -0.08 DWM × Temperature NS -0.07 No-till × Spring 0.06 NS Landscape × Landscape Interactions Clay:sand ratio × Site relief NS -0.71 Clay:sand ratio × Very high STP NS -0.41 Landscape × Weather Interactions Very high STP × Daily rainfall 0.31 NS Site relief × Previous day rainfall 0.17 NS Drain spacing × Temperature -0.10 NS Site relief × Daily rainfall NS 0.53
[a] NS indicates that this factor was not statistically significant at p < 0.05 for the given prediction equation.
Weather Effects on NO3-N and Soluble P Concentrations
Weather effects on NO3-N and soluble P concentrations were strongly driven by season and rainfall. NO3-N concentrations were significantly higher in spring than in other seasons (table 1). Soluble P concentrations were significantly higher in summer than in other seasons (table 1). Recent rainfall events indicated increased concentrations of both NO3-N and soluble P. However, the timing of rainfall impacted these two nutrients differently, as indicated by the magnitude of the factor estimates. A rainfall event of >22 mm (0.9 in.) observed on the day before a sample (factor estimate: 0.72) was a stronger predictor of increased soluble P concentration than a rainfall event of >13 mm (0.5 in.) on the day of the sample (factor estimate: 0.54) (table 1). Cumulative rainfall of >62 mm occurring in the week before the sample was not a significant predictor of soluble P concentration (table 1). However, cumulative weekly rainfall of the same magnitude was a stronger predictor of increased NO3-N concentration (factor estimate: 1.33) than a rainfall event of >22 mm on the day of the sample (factor estimate: 0.87) (table 1). Rainfall occurring the day before a sample was not a significant predictor of NO3-N concentration (table 1). The interaction between “daily rainfall” and “weekly rainfall” was statistically significant for NO3-N concentration (factor estimate: 1.27) and indicated that weekly rainfall was a stronger predictor of increased NO3-N concentration if the observed rainfall was >13 mm on the day of the sample (table 2).
The impact of rainfall on NO3-N and soluble P concentrations varied by season. The response of NO3-N concentration to daily rainfall was less at lower temperatures and during winter (table 2). The response of NO3-N concentration to the previous day’s rainfall was greater at higher temperatures (table 2). During summer months, soluble P concentrations were significantly higher in subsurface drainage (table 1). In addition, the effect of “previous day rainfall” on soluble P concentrations was greater in summer, as shown by the interaction between “summer” and “previous day rainfall” in table 2.
Landscape Effects on NO3-N and Soluble P Concentrations
The site-specific landscape characteristics (STP, soil texture, and slope) significantly affected soluble P concentrations. Fields with STP above the recommended maintenance level (>30 ppm) were observed to have higher concentrations of soluble P in subsurface drainage (table 1). Sites with a higher ratio of clay to sand in the soil profile tended to have lower soluble P concentrations in subsurface drainage. The effect of soil texture on soluble P concentrations was greater when observed with STP > 30 ppm (table 2). Soil texture effects were also greater at sites with high relief (table 2).
NO3-N concentrations were also greater at sites with STP > 30 ppm (table 1). Fields with higher sand content tended to have lower concentrations of NO3-N in subsurface drainage, except when high amounts of fertilizer were applied, as shown by the significant interaction of “sand content × N rate applied” in table 2. Although drain spacing did not show an overall consistent trend, at higher temperatures, sites with narrower drain spacing tended to have greater soluble NO3-N concentrations (table 2).
Interactions between landscape characteristics and rainfall were observed for both NO3-N and soluble P concentrations. The significant interaction of “very high STP × daily rainfall” indicated that the NO3-N concentrations at sites with high STP had a greater increase in NO3-N concentration in response to rainfall than sites with STP below the recommended maintenance level. A significant interaction between “site relief” and “daily rainfall” indicated that rainfall on the day of the sample increased soluble P concentrations by a higher amount at sites with higher relief than at sites with lower relief (table 2). The interaction between “previous day rainfall” and “site relief” for NO3-N suggests a similar effect.
Management Effects on NO3-N and Soluble P Concentrations
A high rate of P fertilizer application was the strongest predictor of elevated soluble P concentrations in subsurface drainage of any factor tested in this study (factor estimate: 1.41) (table 1). This effect indicated that soluble P concentrations increased as the P application rate increased. The significant interaction of “P rate applied × previous day rainfall” was the second strongest predictor of increased soluble P concentrations (factor estimate: 1.34) (table 2). The ability of P application rate to increase soluble P concentration was slightly stronger at sites with high clay:sand ratios (table 2). When P fertilizer was applied during autumn, the soluble P concentrations in subsurface drainage were elevated during winter months, as shown by the significant interaction of “days since P application × winter” in table 2.
The NO3-N concentrations in subsurface drainage were more strongly correlated with fertilizer timing than with fertilizer rate. NO3-N concentrations declined as the amount of time since fertilizer application increased (table 1). The significant interaction of “days since N application × N rate applied” showed that this decline occurred more rapidly following higher rates of fertilizer application (table 2). Fertilizer timing was more strongly correlated with elevated NO3-N concentrations during autumn than during spring (table 2).
Although fertilizer application rate alone was not a significant predictor of NO3-N concentrations, it had significant interactions with other factors. Sites with narrower drain spacing had both a greater increase in NO3-N concentration in response to higher rates of fertilizer application and a greater decline in NO3-N concentration over time than sites with wider drain spacing (table 2). Fertilizer application rate was less correlated to NO3-N concentrations either if higher amounts of rainfall or if warmer temperatures occurred near the time of the sample, as shown by the significant interactions of “N rate applied × daily rainfall” and “N rate applied × temperature” in table 2. This indicates that other factors became stronger drivers of NO3-N loss than fertilizer application rate under high-rainfall or high-temperature conditions. Additionally, fertilizer application rate was less correlated to NO3-N concentration if STP was very high (table 2).
Tillage did not have a significant effect on soluble P unless paired with other factors. No-till management was correlated with increased soluble P concentrations in subsurface drainage only if sites had STP above the agronomic limit (table 2). No-till was correlated with lower soluble P concentrations when sites had higher relief. No-till sites also tended to have lower soluble P concentrations during winter or following a week of high cumulative rainfall (table 2).
Overall, subsurface drainage from continuous no-till sites tended to have lower concentrations of NO3-N. This effect was not as strong during the non-growing season (winter to spring) or at higher temperatures (table 2). Both the timing and rate of N application impacted the effect of tillage on NO3-N concentrations. Following greater applications of N fertilizer, there was less difference between no-till and rotationally tilled sites. However, over time, NO3-N concentrations from no-till sites tended to decrease relative to concentrations in rotationally tilled sites, as shown by the interaction between “no-till” and “days since application” in table 2.
Soybean and wheat crops were correlated with increased soluble P concentrations in subsurface drainage (table 1). The interactions of “soybeans × daily rainfall,” “soybeans × previous day rainfall,” and “wheat × daily rainfall” indicate that soluble P concentrations had a greater increase in response to rainfall during soybean and wheat years compared to corn years (table 2). Concentrations of soluble P in subsurface drainage were greater during summer in wheat years compared to other crop years (table 2).
Fewer significant effects related to soybean and wheat crops were seen for NO3-N than for soluble P concentrations. Wheat crops were correlated with lower concentrations of NO3-N in subsurface drainage (table 1). A significant interaction between soybeans and site relief was observed, indicating that NO3-N concentrations tended to be higher at sites growing soybeans with greater site relief (table 2).
Popcorn crops were not significantly different from corn crops when comparing soluble P concentrations. However, there were two significant interactions when comparing NO3-N concentrations: “popcorn × days since N application” and “popcorn × weekly rainfall.” These two interactions suggest that NO3-N concentrations in subsurface drainage exhibited a faster decline over time following N application in popcorn than in corn crops and that NO3-N concentrations were greater in popcorn than in corn following a period of high weekly rainfall (table 2).
Controlled drainage had a minimal effect on NO3-N and soluble P concentrations relative to the other site-specific factors evaluated in this study. A higher N application rate was more strongly correlated to increased NO3-N concentrations at controlled drainage sites relative to conventionally drained sites (table 2). For soluble P concentrations, a significant interaction was observed between controlled drainage and three factors: temperature, STP, and site relief (table 2). Soluble P concentrations tended to be lower with controlled drainage relative to the concentrations with conventional drainage if temperatures were low, if the site had very high STP, or if the site had higher relief.
Rainfall was a strong predictor of both increased NO3-N and increased soluble P concentration in subsurface drainage. This demonstrates that rainfall acts as a driver of soluble nutrient loss from subsurface drains by moving soluble nutrients through the soil profile. While NO3-N and soluble P concentrations both showed a strong response to rainfall, the difference between their responses to rainfall timing suggests that these two soluble nutrients moved along different pathways through the soil profile. NO3-N concentrations were more likely to be mobilized following longer periods of heavy rainfall. High cumulative weekly rainfall results in high rates of subsurface drainage and thus is likely to cause leaching of NO3-N.
The shorter response of soluble P to rainfall relative to NO3-N indicates that soluble P was more likely to move to subsurface drains via preferential flow pathways. Finer-textured soils, such as those in this study, are more prone to formation of soil macropores, such as earthworm burrows and cracks. This network of macropores in the soil profile moves water rapidly from the soil surface to subsurface drainage systems (Simard et al., 2000; Shipitalo et al., 2004). Movement of water through preferential flow pathways is believed to desorb available P from the soil matrix and increase soluble P loss through subsurface drains (Gächter et al., 2004). The in-creased response of soluble NO3-N concentrations to rainfall on the day of the sample could indicate that soluble NO3-N also moved to subsurface drains via preferential flow pathways but at much smaller rates than soluble P.
The seasonal changes in NO3-N and soluble P concentrations observed in this study could reflect a seasonal transition between baseflow-driven and event-driven subsurface drainage. Subsurface drainage is primarily baseflow-driven during the non-growing season and event-driven during summer (King et al., 2014; Macrae et al., 2007; Lam et al., 2016). The increased temperatures and dry soil conditions prevalent during the summer are likely to cause cracks to form (particularly in the upper layers of the soil profile), promoting greater soluble P loss via preferential flow. Event-based drain discharge has previously been associated with preferential flow and higher concentrations of soluble P in subsurface drainage (Gentry et al., 2007; King et al., 2015a; Williams et al., 2015c). Thus, the concentration of soluble P in subsurface drainage increases as baseflow transitions to event flow in summer. Similarly, a decline in baseflow, and therefore a decline in soil matrix transport of soluble nutrients, would be expected to decrease NO3-N transport to subsurface drains (Kladivko et al., 2004).
The seasonal decline in NO3-N concentrations during summer could also be attributed to a decline in available N because a growing crop is in the field. However, soluble P concentrations do not appear to decrease in the same way in response to a growing crop. This is likely a secondary indicator of event-driven discharge from subsurface drains. P concentrations may decrease in the soil matrix due to P uptake by the growing crop. However, this is masked by the increased soluble P concentrations from event-driven discharge.
Fertilizer Application and Risk of NO3-N and Soluble P Loss
Many sites tested above the maintenance limit of 30 ppm STP according to Tri-State (Ohio, Michigan, and Indiana) fertilizer recommendations for corn and soybeans (Vitosh et al., 1995). Soils at or above the maintenance limit are not expected to respond to additional fertilizer inputs. Soils with higher available P in the subsoil have been reported to have higher concentrations of P in subsurface drainage (McDowell and Sharpley, 2001; Kleinman et al., 2015). The results of this study suggest that it is critical for farmers to adhere to Tri-State fertilizer recommendations to reduce the risk of both NO3-N and soluble P losses to surface waters.
The examination of tillage in relation to other factors demonstrated subtleties in the response of soluble P to tillage that are difficult to determine by looking at tillage alone. An increase in soluble P concentrations in no-till systems was only seen in conjunction with STP above the recommended maintenance limit. This provides evidence of soluble P stratification in the soil profile and transport via preferential flow.
Tillage helps to break up the preferential flow pathways in the soil profile (Cullum, 2009) and mixes P fertilizer into the soil profile. In a 2013 survey, about 70% of Western Lake Erie Basin farmers reported broadcasting P fertilizer, and 36% reported broadcasting P without incorporation (LaBarge and Prochaska, 2014). If fertilizer is broadcast-applied without soil tillage, then P remains on the surface. This results in strat-ification, with much higher concentrations of soil P in the top several inches of the soil profile (Simard et al., 2000). This scenario creates a high risk for P loss from the surface to subsurface drains via preferential flow pathways during rainfall events.
The correlation between no-till management and very high STP levels with higher soluble P concentrations does not suggest that no-till management causes soluble P loss. Rather, it suggests that no-till management could increase the water quality risk from preferential flow based on the preferred fertilizer application method in this watershed. The observed increase in soluble P concentrations during soybean and wheat years provides additional evidence of P loss via preferential flow pathways. Soybean and wheat crops are typically planted without tillage in this watershed, thereby allowing increased macropore formation. These findings underscore the importance of incorporation or subsurface placement of fertilizer under no-till management in the Western Lake Erie Basin.
Impact of Controlled Drainage on NO3-N and Soluble P Concentrations
The interaction effects of controlled drainage with other factors were small. Thus, changes in NO3-N or soluble P concentration due to controlled drainage management are not likely to have a noticeable effect on nutrient loading because load is mainly a function of discharge volume (e.g., Williams et al., 2015b). These findings are in agreement with those of other studies which found that DWM does not greatly impact P or NO3-N concentrations (Wesstrom and Messing, 2007; Sunohara et al., 2015; Williams et al., 2015b).
It is possible that controlled drainage management did not consistently maintain an elevated water table across the sites. For changes in nutrient concentration to occur with implementation of controlled drainage, the water table needs to remain elevated for an extended period to induce redox conditions. Redox conditions promote decreases in NO3-N concentrations through denitrification, but they promote increases in P concentrations through increased P mobility (McDowell et al., 2012). While a more consistent anoxic zone would be beneficial for reduction of NO3-N, it would be detrimental with respect to soluble P concentrations. An inconsistent anoxic zone above the drains could be due to factors not examined in this study, such as unoptimized management of the control structure or increased water movement via lateral and vertical seepage.
It is critical that farmers follow Tri-State fertilizer recommendations and balance crop requirements with fertilizer application rates. Both the timing and rate of fertilizer application were strong predictors of increased NO3-N and soluble P concentrations in subsurface drainage. To mitigate soluble nutrient losses following fertilizer application, farmers should use in-field best management practices (BMPs) to reduce the source of nutrients in the soil profile (e.g., the 4R nutrient management strategy).
The results of this study suggest that a one-size-fits all approach to in-field nutrient transport will not reduce NO3-N and soluble P movement to subsurface drains. Soil matrix flow represents a greater proportion of subsurface drainage than preferential flow and is primarily responsible for NO3-N transport. However, preferential flow can contribute disproportionately to the overall load of soluble P (Stamm et al., 1998; Cullum, 2009). BMPs targeting NO3-N transport to subsurface drains should primarily focus on reducing the soil matrix transport of N to subsurface drains (e.g., use cover crops to scavenge N and increase ET after harvest). BMPs targeting soluble P should primarily focus on reducing P transport via preferential flow (e.g., subsurface placement or incorporation of P fertilizer).
To reduce the risk of downstream water quality impacts in the Western Lake Erie Basin, both the source and transport must be addressed to reduce NO3-N and soluble P losses. When introducing a BMP that could alter the movement of water through the soil profile (e.g., different types of cover crops or no-till), it may be necessary to compensate for potential changes in field hydrology through a stacking of BMPs. Shifts between baseflow and event flow were observed to drive shifts between NO3-N and soluble P loss. Thus, management-driven approaches to reduce transport from one pathway may exacerbate losses from the other. End-of-pipe BMPs (e.g., controlled drainage, bioreactors, and saturated buffers) could be particularly beneficial when stacking BMPs. Further research is needed into how different BMPs alter field hydrology, particularly with respect to preferential flow pathways, to better inform recommendations for managing in-field transport to subsurface drains.
Although controlled drainage did not have a notable effect on NO3-N or soluble P concentration, this practice can still be part of a holistic nutrient management strategy. Controlled drainage has been reported in many studies to provide an overall water quality benefit though discharge and load reductions (e.g., Wesstrom and Messing, 2007; Sunohara et al., 2015; Williams et al., 2015b). Controlled drainage is primarily a physical practice to reduce discharge and subsequently soluble nutrient loads. If a farmer’s primary goal is NO3-N or soluble P concentration reduction, rather than overall load reduction, alternate BMPs (e.g., denitrifying bioreactors or phosphorus removal structures) may be more appropriate. Controlled drainage has strong potential for stacking with other BMPs. Thus, controlled drainage could be implemented to reduce NO3-N and soluble P transport via subsurface drainage while other BMP strategies are implemented to reduce NO3-N and soluble P concentrations.
A multivariate regression approach was used to evaluate the statistical relationships between weather, landscape, and management factors on NO3-N and soluble P concentrations in subsurface drainage for the Western Lake Erie Basin over a study period from 2011 to 2014.
Results of this study strongly suggest that fertilizer management is critical to reducing NO3-N and soluble P concentrations in subsurface drainage waters. The date of fertilizer application was a strong predictor of increased NO3-N concentrations. Fertilizer rate was one of the strongest predictors of soluble P concentrations. Soil test P (STP) was a significant predictor of both soluble P and NO3-N. Increases in NO3-N and soluble P concentrations following fertilizer applications in winter months and/or during soybean years particularly demonstrated the importance of following Tri-State fertilizer recommendations for reducing the risk of soluble nutrient losses to surface waters.
Seasonal variations in NO3-N and soluble P concentrations indicated a change in nutrient dynamics that is tied to transitions between event flow and baseflow in subsurface drainage. The response of soluble P concentrations to rainfall near the time of the sample and interactions between rainfall, tillage, and soluble P fertilizer application suggest movement of soluble P from the soil surface to subsurface drains via preferential flow pathways. The response of NO3-N concentrations to weekly rainfall and season indicate transport to subsurface drains primarily via soil matrix flow.
Reducing nutrient losses from agricultural lands in the Western Lake Erie Basin will require both in-field and edge-of-field approaches. To reduce in-field sources of NO3-N and soluble P, it is critical that farmers:
- Follow Tri-State fertilizer recommendations for application rates based on the current season’s crop.
- Do not apply P fertilizer to soils with STP above the maintenance level.
- Avoid pre-application of N fertilizer before planting.
- Avoid application of P fertilizer before large (>25 mm) rainfall events.
- Perform incorporation (or subsurface placement) of P fertilizer to reduce soluble P transport via preferential flow pathways.
- Stack BMPs to maximize water quality protection and mitigate potential changes in field hydrology, particularly with respect to changes in preferential flow pathways.
This research is part of a regional collaborative project supported by the USDA-NIFA (Award No. 2011-68002-30190, “Cropping Systems Coordinated Agricultural Project: Climate Change, Mitigation, and Adaptation in Corn-Based Cropping Systems,” project website: sustainablecorn.org). The eleven institutions comprising the project team include the following Land Grant universities and the USDA Agricultural Research Service (ARS): Iowa State University, Lincoln University, South Dakota State University, University of Illinois, University of Minnesota, University of Missouri, University of Wisconsin, and the USDA-ARS in Columbus, Ohio.
This research was also funded in part by a USDA-NRCS Conservation Innovation Grant through the Agricultural Drainage Management Coalition (admcoalition.com), an Ohio USDA-NRCS State Conservation Innovation Grant through the Maumee Valley RC&D (Overholt Drainage Education and Research Program, Department of Food, Agricultural and Biological Engineering, Ohio Agricultural Research and Development Center, Ohio State University Extension, the Ohio State University), and was conducted in collaboration with the USDA-ARS Soil Drainage Research Unit. We also thank the cooperating farmers for providing their farms as research and demonstration sites and sharing their farm management information, and Ron Pease, technical writer, for assistance editing the manuscript.
Akaike, H. (1974). A new look at the statistical model identification. IEEE Trans. Auto. Control, 19(6), 716-723. https://doi.org/10.1109/TAC.1974.1100705
Ashraf, M., Loftis, J. C., & Hubbard, K. G. (1997). Application of geostatistics to evaluate partial weather station networks. Agric. Forest Meteorol., 84(3-4), 255-271. https://doi.org/10.1016/S0168-1923(96)02358-1
Beauchamp, K. H. (1987). A history of drainage and drainage methods. In G. Pavelis (Ed.), Farm drainage in the United States: History, status, and prospects (pp. 29-45). Miscellaneous Publ. No. 1455. Washington, DC: USDA Economic Research Service.
Burnham, K. P., & Anderson, D. R. (2004). Multimodel inference: Understanding AIC and BIC in model selection. Sociolog. Methods Res., 33(2), 261-304. https://doi.org/10.1177/0049124104268644
Carpenter, S. R., Caraco, N. F., Correll, D. L., Howarth, R. W., Sharpley, A. N., & Smith, V. H. (1998). Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol. Appl., 8(3), 559-568. https://doi.org/10.1890/1051-0761(1998)008[0559:NPOSWW]2.0.CO;2
Cullum, R. F. (2009). Macropore flow estimations under no-till and till systems. Catena, 78(1), 87-91. https://doi.org/10.1016/j.catena.2009.03.004
Daloglu, I., Cho, K. H., & Scavia, D. (2012). Evaluating causes of trends in long-term dissolved reactive phosphorus loads to Lake Erie. Environ. Sci. Tech., 46(19), 10660-10666. https://doi.org/10.1021/es302315d
Gächter, R., Steingruber, S. M., Reinhardt, M., & Wehrli, B. (2004). Nutrient transfer from soil to surface waters: Differences between nitrate and phosphate. Aquatic Sci., 66(1), 117-122. https://doi.org/10.1007/s00027-003-0661-x
Gentry, L. E., David, M. B., Royer, T. V., Mitchell, C. A., & Starks, K. M. (2007). Phosphorus transport pathways to streams in tile-drained agricultural watersheds. J. Environ. Qual., 36(2), 408-415. https://doi.org/10.2134/jeq2006.0098
Jaynes, D. B., & James, D. E. (2007). The extent of farm drainage in the United States. Soil and Water Conserv. Soc. Intl. Conf. Final Program and Abstracts. Ankeny, IA: Soil and Water Conservation Society.
King, K. W., Fausey, N. R., & Williams, M. R. (2014). Effect of subsurface drainage on streamflow in an agricultural headwater watershed. J. Hydrol., 519, 438-445. https://doi.org/10.1016/j.jhydrol.2014.07.035
King, K. W., Williams, M. R., & Fausey, N. R. (2015a). Contributions of systematic tile drainage to watershed-scale phosphorus transport. J. Environ. Qual., 44(2), 486-494. https://doi.org/10.2134/jeq2014.04.0149
King, K. W., Williams, M. R., Macrae, M. L., Fausey, N. R., Frankenberger, J., Smith, D. R., ... Brown, L. C. (2015b). Phosphorus transport in agricultural subsurface drainage: A review. J. Environ. Qual., 44(2), 467-485. https://doi.org/10.2134/jeq2014.04.0163
Kladivko, E. J., Frankenberger, J. R., Jaynes, D. B., Meek, D. W., Jenkinson, B. J., & Fausey, N. R. (2004). Nitrate leaching to subsurface drains as affected by drain spacing and changes in crop production system. J. Environ. Qual., 33(5), 1803-1813. https://doi.org/10.2134/jeq2004.1803
Kleinman, P. J. A., Church, C., Saporito, L. S., McGrath, J. M., Reiter, M. S., Allen, A. L., ... Joern, B. C. (2015). Phosphorus leaching from agricultural soils of the Delmarva Peninsula, USA. J. Environ. Qual., 44(2), 524-534. https://doi.org/10.2134/jeq2014.07.0301
LaBarge, G. A., & Prochaska, S. C. (2014). Soil testing and nutrient application practices of Ohio agronomy retailers. Marion, OH: Ohio State University Extension.
Lam, W. V., Macrae, M. L., English, M. C., O’Halloran, I. P., Plach, J. M., & Wang, Y. (2016). Seasonal and event-based drivers of runoff and phosphorus export through agricultural tile drains under sandy loam soil in a cool temperate region. Hydrol. Proc., 30(15), 2644-2656. https://doi.org/10.1002/hyp.10871
Loftis, J. C., & Ward, R. C. (1980). Water quality monitoring: Some practical sampling frequency considerations. Environ. Mgmt., 4(6), 521-526. https://doi.org/10.1007/bf01876889
Macrae, M. L., English, M. C., Schiff, S. L., & Stone, M. (2007). Intra-annual variability in the contribution of tile drains to basin discharge and phosphorus export in a first-order agricultural catchment. Agric. Water Mgmt., 92(3), 171-182. https://doi.org/10.1016/j.agwat.2007.05.015
McDowell, R. W., & Sharpley, A. N. (2001). Approximating phosphorus release from soils to surface runoff and subsurface drainage. J. Environ. Qual., 30(2), 508-520. https://doi.org/10.2134/jeq2001.302508x
McDowell, R. W., Gongol, C., & Woodward, B. (2012). Potential for controlled drainage to decrease nitrogen and phosphorus losses to Waituna Lagoon. Invercargill, New Zealand: Environment Southland.
Menne, M. J., Durre, I., Vose, R. S., Gleason, B. E., & Houston, T. G. (2012). An overview of the Global Historical Climatology Network - Daily database. J. Atmos. Oceanic Tech., 29(7), 897-910. https://doi.org/10.1175/jtech-d-11-00103.1
Moog, D. B., & Whiting, P. J. (2002). Climatic and agricultural factors in nutrient exports from two watersheds in Ohio. J. Environ. Qual., 31(1), 72-83. https://doi.org/10.2134/jeq2002.7200
Murphy, J., & Riley, J. P. (1962). A modified single-solution method for the determination of phosphate in natural waters. Anal. Chim. Acta, 27, 31-36. https://doi.org/10.1016/S0003-2670(00)88444-5
NCDC. (2016). NOAA’s Gridded Climate Divisional Dataset (CLIMDIV) [U.S. Climate Division 3301]. Asheville, NC: National Climatic Data Center.
OLEPTF. (2010). Ohio Lake Erie Phosphorus Task Force Final Report. Columbus, OH: Ohio Environmental Protection Agency. Retrieved from http://www.epa.state.oh.us/portals/35/lakeerie/ptaskforce/Task_Force_Final_Report_April_2010.pdf
Pease, L. A. (2016). Characterization of agricultural subsurface drainage water quality and controlled drainage in the Western Lake Erie Basin. PhD diss. Columbus, OH: Ohio State University, Department of Food, Agricultural, and Biological Engineering.
Rucinski, D. K., Beletsky, D., DePinto, J. V., Schwab, D. J., & Scavia, D. (2010). A simple one-dimensional, climate-based dissolved oxygen model for the central basin of Lake Erie. J. Great Lakes Res., 36(3), 465-476. https://doi.org/10.1016/j.jglr.2010.06.002
Sands, G. R., Song, I., Busman, L. M., & Hansen, B. J. (2008). The effects of subsurface drainage depth and intensity on nitrate loads in the northern cornbelt. Trans. ASABE, 51(3), 937-946. https://doi.org/10.13031/2013.24532
SAS. (2013). JMP Pro (ver. 11.0.0). Cary, NC: SAS Institute, Inc.
Selman, M., Greenhalgh, S., Diaz, R., & Sugg, Z. (2008). Eutrophication and hypoxia in coastal areas: A global assessment of the state of knowledge. WRI Policy Note. Washington, DC: World Resources Institute.
Shipitalo, M. J., Nuutinen, V., & Butt, K. R. (2004). Interaction of earthworm burrows and cracks in a clayey, subsurface-drained soil. Appl. Soil Ecol., 26(3), 209-217. https://doi.org/10.1016/j.apsoil.2004.01.004
Simard, R. R., Beauchemin, S., & Haygarth, P. M. (2000). Potential for preferential pathways of phosphorus transport. J. Environ. Qual., 29(1), 97-105. https://doi.org/10.2134/jeq2000.00472425002900010012x
Smith, D. R., King, K. W., Johnson, L., Francesconi, W., Richards, P., Baker, D., & Sharpley, A. N. (2015). Surface runoff and tile drainage transport of phosphorus in the midwestern United States. J. Environ. Qual., 44(2), 495-502. https://doi.org/10.2134/jeq2014.04.0176
Sprague, L. A., & Gronberg, J. A. M. (2012). Relating management practices and nutrient export in agricultural watersheds of the United States. J. Environ. Qual., 41(6), 1939-1950. https://doi.org/10.2134/jeq2012.0073
Stamm, C., Fluhler, H., Gachter, R., Leuenberger, J., & Wunderli, H. (1998). Preferential transport of phosphorus in drained grassland soils. J. Environ. Qual., 27(3), 515-522. https://doi.org/10.2134/jeq1998.00472425002700030006x
Sunohara, M. D., Gottschall, N., Wilkes, G., Craiovan, E., Topp, E., Que, Z., ... Lapen, D. R. (2015). Long-term observations of nitrogen and phosphorus export in paired agricultural watersheds under controlled and conventional tile drainage. J. Environ. Qual., 44(5), 1589-1604. https://doi.org/10.2134/jeq2015.01.0008
Turner, R. E., & Rabalais, N. N. (2003). Linking landscape and water quality in the Mississippi River basin for 200 years. BioScience, 53(6), 563-572. https://doi.org/10.1641/0006-3568(2003)053[0563:LLAWQI]2.0.CO;2
USDA-NRCS. (2015). Tillage study results - WLEB. Washington, DC: USDA Natural Resources Conservation Service. Retrieved from http://www.nrcs.usda.gov/wps/portal/nrcs/detail/oh/technical/landuse/cropland/?cid=nrcs144p2_029581
USEPA. (1983). Methods for chemical analysis of water and wastes. Washington, DC: U.S. Environmental Protection Agency.
Vitosh, M. L., Johnson, J. W., & Mengel, D. B. (1995). Tri-State fertilizer recommendations for corn, soybeans, wheat, and alfalfa. Extension Bulletin E-2567. East Lansing, MI: Michigan State University Extension.
Wesstrom, I., & Messing, I. (2007). Effects of controlled drainage on N and P losses and N dynamics in a loamy sand with spring crops. Agric. Water Mgmt., 87(3), 229-240. https://doi.org/10.1016/j.agwat.2006.07.005
Williams, M. R., King, K. W., & Fausey, N. R. (2015a). Contribution of tile drains to basin discharge and nitrogen export in a headwater agricultural watershed. Agric. Water Mgmt., 158(3), 42-50. https://doi.org/10.1016/j.agwat.2015.04.009
Williams, M. R., King, K. W., & Fausey, N. R. (2015b). Drainage water management effects on tile discharge and water quality. Agric. Water Mgmt., 148, 43-51. https://doi.org/10.1016/j.agwat.2014.09.017
Williams, M. R., King, K. W., Macrae, M. L., Ford, W., Van Esbroeck, C., Brunke, R. I., ... Schiff, S. L. (2015c). Uncertainty in nutrient loads from tile-drained landscapes: Effect of sampling frequency, calculation algorithm, and compositing strategy. J. Hydrol., 530, 306-316. https://doi.org/10.1016/j.jhydrol.2015.09.060