Dissolved organic carbon loading from the field to watershed scale in tile-drained landscapes

Agricultural Water Management 192 (2017) 159–169

Contents lists available at ScienceDirect

Agricultural Water Management journal homepage: www.elsevier.com/locate/agwat

Research paper

Dissolved organic carbon loading from the field to watershed scale in tile-drained landscapes Mark R. Williams a,∗ , Kevin W. King b , Norman R. Fausey b a b

USDA Agricultural Research Service, National Soil Erosion Research Laboratory, 275 South Russell Street, West Lafayette, IN 47907, United States USDA Agricultural Research Service, Soil Drainage Research Unit, 590 Woody Hayes Drive, Columbus, OH 43210, United States

a r t i c l e

i n f o

Article history: Received 23 January 2017 Received in revised form 23 May 2017 Accepted 12 July 2017 Keywords: Subsurface tile drain Headwater Drainage water management Water quality Nutrient transport Subsurface hydrology

a b s t r a c t Subsurface tile drains influence watershed fluxes of nitrogen, phosphorus, and pesticides, but few studies have examined the role of subsurface tile drains and drainage water management practices on watershed dissolved organic carbon (DOC) export. The objective of this study was therefore to quantify the contribution of subsurface tile drains to watershed DOC export and to evaluate the effect of drainage water management of DOC concentrations and loads in tile-drained fields. Discharge and DOC concentration were measured at the outlet of an agricultural headwater watershed (3.9 km2 ) in Ohio, USA and all of the subsurface tile drains (6 total) within the watershed over an 8-year period. Results showed that DOC concentration in both subsurface tile drains and stream water were highly variable (0.1–44.4 mg L−1 ), with mean DOC concentrations ranging from 5.7 to 8.2 mg L−1 . Intra-annual variability in subsurface tile drain and watershed hydrology yielded seasonal differences in DOC loading. Over the study period, 81.7% and 92.4% of watershed and subsurface tile drain DOC loading, respectively, occurred during 20% of the time, typically during winter and spring high flow events. Mean annual DOC loading from the drainage network was 19.6 kg ha−1 , while mean annual DOC loading at the watershed outlet was 43.9 kg ha−1 . On average, subsurface tile drainage comprised 33% of monthly watershed DOC export (<1–82%). Implementing drainage water management at one of the subsurface tile drains decreased discharge (179 mm; 22%) and DOC loading (6.8 kg ha−1 ; 26%) compared to an adjacent free draining subsurface tile drain. Findings from this study demonstrate the utility of simultaneously monitoring solute fluxes from both field and watershed scales, and indicate that subsurface tile drains are a significant source of DOC to headwater agricultural streams. Further, results suggest that drainage water management can significantly decrease DOC losses from tile-drained fields. Published by Elsevier B.V.

1. Introduction Dissolved organic carbon (DOC) is essential to the functioning of aquatic ecosystems and represents a large flux of organic matter from watersheds (e.g., Mulholland, 1997). DOC serves as an energy source for heterotrophs, regulates the cycling of inorganic nutrients, and influences light and temperature regimes (Stanley et al., 2012). It also affects the transport and bioavailability of organic pollutants and heavy metals, which pose a risk to downstream drinking water quality (Ledesma et al., 2012). While the dynamics and biogeochemistry of DOC have been widely studied in upland forested watersheds, few studies have examined DOC fluxes in artificially drained agricultural watersheds commonly found across the Mid-

∗ Corresponding author. E-mail addresses: [email protected] (M.R. Williams), [email protected] (K.W. King), [email protected] (N.R. Fausey). http://dx.doi.org/10.1016/j.agwat.2017.07.008 0378-3774/Published by Elsevier B.V.

western U.S., eastern Canada, and northern Europe. Subsurface tile drainage alters the dominant hydrologic flow pathways in these watersheds (e.g., King et al., 2014), which has been shown to influence fluxes of nitrogen (Tomer et al., 2003; Kennedy et al., 2012; Williams et al., 2015), phosphorus (Macrae et al., 2007; King et al., 2015), and pesticides (Kladivko et al., 2001; Stone and Wilson, 2006). As such, subsurface tile drainage likely also represents an important contributor to the overall DOC flux from agricultural watersheds (Royer and David, 2005; Dalzell et al., 2007; Ruark et al., 2009; Warrner et al., 2009). Research at both field and watershed scales has shown the potential importance of subsurface tile drains on stream DOC export. Field studies examining the effect of specific factors such as vegetation type (Aitkenhead and McDowell, 2000) and drainage intensity (Dalzell et al., 2011) on DOC fluxes have suggested that single tile drain outlets can contribute up to 48 kg ha−1 of DOC to receiving streams and agricultural drainage ditches (e.g., Ruark

160

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

et al., 2009). At the watershed scale, studies have indicated that subsurface tile drainage results in substantial intra-annual variation in stream water DOC concentration compared to undrained watersheds where wetlands persist (Hinton et al., 1997; Dalzell et al., 2007). The magnitude, timing, and composition of DOC in artificially drained watersheds has also been shown to reflect the quick transfer of DOC via subsurface drainage networks to the stream channel (Vidon et al., 2008). Separating watershed DOC load between subsurface tile drains and other potential sources (i.e., surface runoff, in-stream processes) has been a limitation in previous studies of DOC export often due to the number of tile drain outlets within these watersheds. Indeed, DOC fluxes in tile-drained watersheds ranging from 22 to 480 km2 have been examined, but these watersheds likely contain tens to hundreds of tile drain outlets which requires extrapolation of field-scale DOC loads to the watershed scale (e.g., Royer and David, 2005; Warrner et al., 2009). Measuring nutrient export in small headwater watersheds (<5 km2 ) with fewer tile drain outlets, however, has been a useful approach for understanding the relationship between subsurface drainage and watershed nutrient fluxes (e.g., Macrae et al., 2007; King et al., 2015; Williams et al., 2015). Understanding nutrient and carbon transport in small headwater watersheds also has broad implications for biogeochemical cycling in downstream (higher order) systems (e.g., Peterson et al., 2001; Dalzell et al., 2011). Thus, in this study, we simultaneously measure DOC export at the outlet of a small agricultural headwater watershed and all of the subsurface tile drains (6 total) in the watershed to assess the integrated effects of subsurface tile drainage on the magnitude and timing of watershed DOC transport. Managing carbon in agricultural landscapes has become of greater interest in recent years (Stanley et al., 2012), as national and global initiatives are promoting soil health, which includes increasing soil organic matter (NRCS, 2003) and carbon sequestration in agricultural soils. Reduced tillage practices, planting cover crops, and using animal manures have been shown to increase soil organic carbon in agricultural fields (e.g., Lal, 2004), but these increases may be offset in tile-drained fields due to increased DOC losses through the subsurface drainage network. In areas of northern Europe where peatlands have been historically surface-drained to increase the area of land suitable for agriculture or to allow peatcutting for fuel, drain blocking (i.e., creating dams at intervals along the length of the surface drain) has been used to decrease DOC fluxes (Wallage et al., 2006; Armstrong et al., 2010). This suggests that drainage water management, the practice of seasonally adjusting the outlet elevation of the subsurface drainage system through installation of a water control structure (Skaggs et al., 2012), may be an effective practice in tile-drained landscapes to reduce DOC delivery from tile drains to streams. In this study, we measure discharge and DOC concentration at the outlet of an agricultural headwater watershed in central Ohio, USA and all of the subsurface tile drains within the watershed over an 8 year period (2005–2012) to better understand the relationship between field and watershed scale DOC export. Our objectives were to (1) determine the proportion of watershed DOC export that is derived from subsurface tile drainage in an agricultural headwater watershed; and (2) quantify the effect of drainage water management on DOC concentrations and loads in tile-drained fields. Since subsurface tile drainage can have a large influence on hydrologic flow paths in headwater watersheds, we hypothesized that DOC loading from subsurface tile drainage would be a significant contributor to the overall watershed DOC flux. We also hypothesized that implementation of drainage water management would decrease DOC loading from tile-drained fields compared to conventional free drainage.

Table 1 Crop production management including tillage, crop, nutrient source, and nutrient rate, for fields B2 and B4. Management in B2 and B4 is representative of the prevailing practices that were found throughout Watershed B during the study period. Year

Date

Operationa

2004

May 11 May 13

November 13

Tillage Corn planting Fertilizer application; 12-15-20 (26.9 kg N ha−1 ; 14.7 kg P ha−1 ) Fertilizer application; 28-0-0 (167.3 kg N ha−1 ) Harvest

2005

May 7 October 5

Soybean planting Harvest

2006

April 30 May 1

Tillage Corn planting Fertilizer application; 10-34-0 (82.1 kg N ha−1 ; 48.7 kg P ha−1 ) Fertilizer application; 28-0-0 (167.3 kg N ha−1 ) Harvest

June 13

June 20 October 27 2007

2008

May 9 October 10 October 16 October 17

Soybean planting Harvest Fertilizer application; Chicken litter (456 kg N ha−1 ; 117.4 kg P ha−1 ) Tillage

May 7 October 2

Soybean planting Harvest

Adapted from King et al. (2016). a Fertilizer listed as nitrogen-phosphorus-potassium (N-P-K).

2. Materials and methods 2.1. Experimental watershed Our study took place in Watershed B (3.9 km2 ), which is located in central Ohio, USA (40◦ 12 41.83 N, 82◦ 49 31.48 W) and is a subwatershed of the Upper Big Walnut Creek (492 km2 ) (Fig. 1). The Upper Big Walnut Creek watershed is an USDA Agricultural Research Service benchmark watershed and has been studied as part of the Conservation Effects Assessment Project (CEAP) since 2004 (Mausbach and Dedrick, 2004). The majority of headwater streams in the Upper Big Walnut Creek watershed have been classified as impaired due to nutrient enrichment, pathogens, and habitat degradation stemming from agricultural practices (Ohio EPA, 2004). Crop production agriculture consisting primarily of corn (Zea mays L.)-soybean [Glycine max (L.) Merr.] crop rotations comprises 73% of the land use in Watershed B, with the remainder of the watershed consisting of urban/farmstead (21%) and woodland (6%) land uses. Typical crop production management for agricultural fields in Watershed B is shown in Table 1. The topography of Watershed B is relatively flat, with slopes less than 1%. Soils throughout the watershed consist of a somewhat poorly drained Bennington silt loam (Fine, illitic, mesic Aeric Epiaqualfs) and a very poorly drained Pewamo clay loam (Fine, mixed, active, mesic Typic Argiaquolls) (Table 2). These soils typically contain 3–4% organic matter in surface horizons, with organic matter content decreasing with depth (USDA SSURGO, 2017). Soil saturated hydraulic conductivity ranges between 33 mm h−1 at the surface and 3–10 mm h−1 at a depth of 1.0 m (USDA SSURGO, 2017). Based on existing subsurface drainage maps for the watershed and conversations with land owners, it is estimated that 80% of Watershed B is systematically tile-drained (Fig. 1). Subsurface tile laterals are clay or perforated plastic, generally spaced 15 m apart, and buried at a depth of 0.9–1.0 m. The estimated average age of the subsurface drainage network is >50 yr. Surface inlets, one located in a grassed waterway and one in a roadside ditch, are connected to

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

161

Fig. 1. Map of surface and subsurface contributing areas and sampling locations. Sites B2, B3, B4, B5, B6, and B8 are subsurface tile drain outlets; site B1 is the watershed outlet. Table 2 Sampling site characteristics for all monitoring locations within Watershed B including estimated drainage area, average land slope, and soil series. Site

B1 B2 B3 B4 B5 B6 B8

Descriptiona

Watershed outlet Field tile County main Field tile Tile main Tile main Field tile

Drainage area (ha)

389 14 211 15 22 49 8

Slope (m/m)

0.009 0.009 0.007 0.008 0.008 0.010 0.009

Soil seriesb

Bennington (%)

Pewamo (%)

52.9 71.9 37.3 52.2 40.8 43.6 86.3

46.2 28.1 62.8 47.8 59.2 53.2 13.7

Adapted from King et al. (2014). a Tile drain diameter: Field tile (0.2 m); Tile main (0.4 m); and County main (0.6 m). b Centerburg silt loam is also found within B1 (0.9%) and B6 (3.2%).

the subsurface drainage network in B3 and B5, respectively. A total of 7 subsurface tile drain outlets discharge into the drainage ditch in Watershed B. With the exception of one subsurface tile drain outlet that drained approximately 7 ha, all subsurface tile drains were instrumented (Fig. 1). This one tile outlet was not functional for the first 4 years of the study (2005–2009) and resource limitations prevented instrumentation once its functionality was restored. 2.2. Discharge and water quality Stream discharge at the outlet of Watershed B was measured from 2005 through 2012 with a 2.4 m Parshall flume, which was equipped with an Isco (Teledyne Isco, Lincoln, NE) 4230 Bubbler Flow Meter and an Isco 2150 Area Velocity Sensor. For each edgeof-field tile drain, compound weir inserts (Thel-Mar, LLC, Brevard,

NC) were installed at the tile outlet. The existing 20-cm diameter subsurface tile drain outlet pipes were cut and fitted with 30-cm diameter pipes that accommodate the compound weir. For the larger 40- and 60-cm diameter drainage mains, an H-flume was fitted to the end of the tile outlet and served as a control volume. Each of the tile drain outlets was also instrumented with an Isco 4230 Bubbler Flow Meter and an Isco 2150 Area Velocity Sensor. Discharge from the watershed outlet and all subsurface tile drains was measured continuously at 10–30 min intervals. Precipitation was measured using an Isco 674 tipping bucket rain gauge and a standard rain gauge (NovaLynx 260-2510), which were located near site B4. Isco 6712 portable samplers were installed at each subsurface tile drain outlet and at the outlet of Watershed B to collect water samples. Water samples were collected every 6 h and 4 aliquots

162

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

were placed in each bottle to comprise a daily sample. Upon return to the laboratory, daily samples were composited into a weekly sample. All water samples were immediately filtered (0.45 ␮m) and refrigerated (4 ◦ C). DOC concentration was determined by heatedpersulfate oxidation using an Aurora 1030W Total Organic Carbon Analyzer (OI Analytical, College Station, TX) with in-line sample acidification and sparging. DOC concentrations were measured at each of the subsurface tile drains from January 2005 through March 2011, while DOC concentrations at the watershed outlet were measured from January 2005 through December 2012. 2.3. Drainage water management Two adjacent subsurface tile drain outlets, B2 and B4, on the south side of the drainage ditch in Watershed B were selected to investigate the effect of drainage water management on DOC concentration and load. Both subsurface tile drains were located within the same field and were managed according to the same agricultural management practices (e.g., tillage, nutrient management) (Table 1). Drainage water management is the practice of seasonally adjusting the outlet elevation of the drainage system through installation of a water control structure, which is typically comprised of stackable boards or stop logs (Skaggs et al., 2012). The outlet elevation can be set at any level between the ground surface and the drainage depth (0.9 m for B2 and B4). In 2009, a water control structure (Agri Drain Co., Adair, IA) was installed at B4 to manage the elevation of the subsurface tile drain outlet. The goal was to restrict the drainage for as much time as possible, while maintaining adequate drainage for spring and fall field operations and for root zone aeration. Thus, the outlet at B4 was lowered a few weeks before the beginning of spring or fall fieldwork and raised to 45 cm below the ground surface shortly after the spring or fall field work was completed. Drainage water management was implemented at site B4 from January 2009 through the end of the study. B2 was free-draining for the entire study period and served as the control. 2.4. Data analysis and statistics DOC load from each of the subsurface tile drain outlets and the outlet of Watershed B was determined by first using linear interpolation to estimate daily DOC concentration from the weekly concentration dataset. Interpolated DOC concentration was then multiplied by daily discharge to calculate daily DOC load. Daily DOC load was subsequently aggregated to monthly and annual DOC loads for each of the monitoring locations. To quantify the relationship between discharge and DOC concentration, the weekly DOC concentration dataset for each monitoring location was regressed against average weekly discharge. The contribution of subsurface tile drains to watershed DOC export was determined by first summing monthly tile drain DOC loads. Summed tile drain DOC load was then compared to monthly watershed DOC load using linear regression. The effect of season on subsurface tile drain and watershed discharge, DOC concentration, and DOC load was also evaluated. Monthly data were first separated into four seasons: winter (January–March), spring (April–June), summer (July–September), and fall (October–December). Differences among seasons were assessed using a Kruskal-Wallis test. When significant, pairwise comparisons were completed using Dunn’s test with a Bonferroni correction. All analyses were completed in R statistical software (R Foundation for Statistical Computing, 2011) and a probability of 0.05 was used to evaluate statistical significance. To determine the effect of drainage water management on DOC concentration and load at B4, tile discharge and DOC data were analyzed using a before-after control-impact (BACI) study design. Data

were classified as being from either before (2005–2008) or after (2009–2011) the implementation of drainage water management at B4. Site B2 was free-draining throughout the entire study period and served as the control, whereas B4 was considered the impact site due the installation of the control structure and subsequent treatment (i.e., management of the tile drain outlet elevation). The strength of the BACI design lies in the assumption that any changes over time in the impact site, unrelated to the treatment, are controlled for by these same changes over time in the control site. Thus, a significant time (before-after) × treatment (control-impact) interaction indicates that the experimental treatment truly has an effect of the impact site (Smith, 2002). BACI analysis was completed using a repeated measures generalized linear mixed model in R statistical software. 3. Results 3.1. Precipitation and discharge Annual precipitation measured at the rain gauge in Watershed B varied from 773 mm in 2012 to 1,239 mm in 2011 (Fig. 2). Over the 8-yr record, mean annual precipitation was 1004 mm, which was slightly greater than the 30-yr average rainfall (985 mm) recorded at a nearby gauge in the southwest portion of the Upper Big Walnut Creek watershed (NCDC, 2016). February tended to be the driest month (45 mm) while June was the wettest month (122 mm), with average monthly precipitation equaling 84 mm. On average, there were 42 precipitation events per year that measured >6.4 mm. Mean annual discharge at the outlet of Watershed B (i.e., B1) was 508 mm, but ranged from 310 to 767 mm (Fig. 2). Annual watershed discharge tended to follow patterns in precipitation and discharge to precipitation ratios were between 39 and 62% (mean = 50%). Summed subsurface tile discharge (i.e., total discharge from all of the subsurface tile drains in Watershed B) ranged from 140 to 564 mm annually and averaged 283 mm (Fig. 2). Comparing annual watershed discharge and annual summed subsurface tile discharge indicated that 56% of watershed discharge originated from the subsurface drainage network, with annual contributions ranging from 37 to 74%. The difference between watershed discharge and summed tile discharge was representative of other sources of water entering the stream including surface runoff, direct precipitation into the stream channel, and groundwater. Collectively, these other sources of water accounted for 225 mm of watershed discharge annually (88–383 mm). Seasonally, both watershed discharge and summed subsurface tile discharge were significantly greater during the winter compared to the spring, summer, and fall (Table 3). Monthly contributions of summed subsurface tile discharge to watershed discharge varied from 10 to 88% (mean = 47%). In general, summed subsurface tile discharge accounted for a greater fraction of watershed discharge during the winter, spring, and fall (50.9–59.8%) compared to the summer (31.9%) (Table 3). Contributions of summed surface tile discharge were also more variable during the summer compared to the other seasons. 3.2. DOC concentration DOC concentration was similar among the subsurface tile drains monitored in Watershed B (Fig. 3). Subsurface tile drain DOC concentrations were often between 2.6 and 10.2 mg L−1 ; however, DOC concentration measured in subsurface discharge ranged from 0.1 to 37.2 mg L−1 over the study period. Mean DOC concentration for individual subsurface tile drain outlets was between 5.7 and 8.2 mg L−1 . DOC concentrations measured in subsurface tile drains were similar to measured DOC concentration at the watershed out-

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

163

Fig. 2. Precipitation and discharge in Watershed B. (A) Daily discharge from the watershed outlet (B1) and all of the subsurface tile drains (B2, B3, B4, B5, B6, and B8). (B) Annual precipitation, watershed discharge, and subsurface tile drain discharge.

Table 3 Seasonal discharge, DOC concentration, and DOC load from the watershed outlet and subsurface tile drainage. Data for subsurface tile drains and subsurface tile drainage to watershed ratios are from 2005 to 2010, while watershed data are from 2005 to 2012. Values in the same row with different letters are significantly different at p < 0.05. Values shown are mean ± one standard deviation. Winter

Spring

Summer

Fall

Discharge Watershed discharge (mm) Subsurface tile discharge (mm) Tile drainage/watershed (%)

119.5 ± 71.8 a 114.6 ± 35.8 a 59.8 ± 13.8 a

132.6 ± 68.0 a 72.8 ± 52.7 a 50.9 ± 15.2 a

35.7 ± 30.9 b 14.3 ± 17.2 b 31.9 ± 25.5 a

137.1 ± 71.1 a 75.6 ± 50.7 a 56.0 ± 17.0 a

DOC concentration Watershed (mg/L) Subsurface tile drainage (mg/L)

7.6 ± 1.7 a 6.9 ± 1.4 b

9.6 ± 3.7 a 8.6 ± 3.0 a

9.7 ± 2.0 a 9.6 ± 4.3 a

8.3 ± 1.9 a 7.7 ± 2.6 ab

DOC load Watershed (kg ha−1 ) Subsurface tile drainage (kg ha−1 ) Tile drainage/watershed (%)

16.0 ± 7.9 a 8.7 ± 3.1 a 55.1 ± 16.7 a

12.7 ± 7.5 a 5.5 ± 3.4 a 46.3 ± 18.9 a

3.5 ± 3.5 b 1.1 ± 1.3 b 29.0 ± 33.2 a

11.7 ± 7.0 a 4.7 ± 2.3 ab 51.9 ± 17.3 a

let (Fig. 3). Mean DOC concentration at site B1 was 7.4 mg L−1 and varied between 0.1 and 44.4 mg L−1 .

Patterns in monthly flow-weighted DOC concentration in subsurface tile drains were variable (Fig. 3). On average, monthly

164

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

and the watershed outlet, DOC concentration was more variable during low flow periods typically observed during the summer compared to high flow periods typically observed during the winter and spring. DOC concentration exhibited a significant positive linear relationship with discharge for all of the subsurface tile drains as well as the watershed outlet (Fig. 4). Despite statistical significance, discharge was not a strong predictor of DOC concentration, as R2 values were between 0.03 and 0.08. 3.3. DOC loading

Fig. 3. DOC concentration. (A) Box-and-whisker plot of DOC concentrations from each sampling location. Circles represent the 5th and 95th percentiles, whiskers represent the 10th and 90th percentiles, the lower and upper edges of the boxes represent the 25th and 75th percentiles, the horizontal line inside of the box represents the median and the × represents the mean. (B) Monthly flow-weighed mean DOC concentration at B1 (watershed outlet). (C) Monthly flow-weighted mean DOC concentration from individual subsurface tile drain outlets (B2 [ ]; B3 [ ]; B4 [ ]; B5 [ ]; B6 [ ]; and B8 [ ]). Average monthly subsurface tile drainage DOC concentration also shown.

flow-weighted mean DOC concentration in subsurface tile drains tended to increase from 6.6 mg L−1 in January to 10.5 mg L−1 in August and September before decreasing back to 6.0 mg L−1 in December. Seasonally, flow-weighted mean DOC concentration was significantly greater in subsurface tile drainage during the spring and summer compared to the winter (Table 3). In contrast to DOC concentration in subsurface tile drains, monthly flowweighted mean DOC concentration at the watershed outlet was less variable, with DOC concentrations remaining close to 8.0 mg L−1 throughout the entire year (Fig. 3). Indeed, flow-weighted mean DOC concentration at the watershed outlet was not significantly different among seasons (Table 3). The relationship between weekly average discharge and DOC concentration is shown in Fig. 4. For both subsurface tile drains

Annual DOC loading from individual subsurface tile drains and the watershed outlet tended to follow trends in precipitation and discharge, with greater loads observed during wet years compared to dry years (Table 4; Fig. 2). DOC loading from individual subsurface tile drains ranged from 6.6 to 66.0 kg ha−1 yr−1 . B5 had the greatest mean annual load (46.9 kg ha−1 ) compared to the other subsurface tile drains (14.6–31.2 kg ha−1 ). Annual summed subsurface tile drain DOC load was between 15.5 and 26.1 kg ha−1 , and averaged 19.6 kg ha−1 (Table 4). Watershed DOC loading varied between 24.1 and 71.7 kg ha−1 yr−1 , with a mean annual DOC load of 43.9 kg ha−1 . Seasonally, summed subsurface tile drain DOC load was significantly greater in the winter and spring compared to summer (Table 3). Mean monthly DOC loading from summed subsurface tile drainage ranged from 0.3 kg ha−1 in August to 4.3 kg ha−1 in March (Fig. 5). Similarly, watershed DOC loading was significantly greater during the winter, spring, and fall compared to the summer, with mean monthly DOC loading between 1.0 (July) and 6.2 kg ha−1 (March) (Table 3; Fig. 5). Monthly summed subsurface tile drain DOC load, on average, comprised 33% of the monthly watershed DOC load, with values ranging from <1 to 82% (Fig. 5). Subsurface tile drain DOC load generally comprised a greater proportion of watershed DOC load during the winter, spring, and fall (46.3–55.1%) compared to the summer (29.0%) (Table 3). DOC loads from individual subsurface tile drains and the watershed outlet were generally less than 0.5 kg d−1 , except during high flow events. Daily DOC loads from all monitoring locations were ranked over the study period to better understand the timing of DOC loading. Over the 8 year study period, 81.7% of watershed DOC loading occurred during 20% of the time (Fig. 6). For subsurface tile drains, between 88.4 and 96.9% (mean = 92.4%) of DOC loading occurred during 20% of the time (Fig. 6). These results highlight the importance of high flow events on DOC delivery in tile-drained landscapes. 3.4. Drainage water management The effect of drainage water management on weekly subsurface tile drain discharge, DOC concentration, and DOC load was evaluated using subsurface tile drains B2 and B4. Subsurface tile drain discharge did not vary significantly between sites or between periods (i.e., 2005–2008 vs. 2009–2011); however, a significant time × treatment interaction was found (Table 5). This suggests that drainage water management significantly impacted subsurface tile drain discharge at B4. The linear relationship between discharge at B2 and B4 from 2005 through 2008 was used to provide an estimate of discharge for B4 over the 2009–2011 period as if drainage water management was not implemented. Since B2 and B4 were draining the same field with like management practices (Table 1), we assumed that the linear relationship established between sites would have remained consistent if drainage water management had not been implemented at site B4. Comparing observed and predicted values for discharge at B4 suggests that drainage water management decreased discharge by 179 mm (22%) between January 2009 and March 2011.

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

165

Fig. 4. Relationship between average weekly discharge and DOC concentration. (A) Site B1 (watershed outlet) and (B) site B2 (subsurface tile drain) shown as examples.

Fig. 5. Monthly DOC loads. (A) Monthly summed subsurface tile drainage DOC load regressed against monthly watershed DOC load. (B) Mean monthly summed subsurface tile drain and watershed outlet DOC load. Error bars represent one standard deviation.

Similar to subsurface tile drain discharge, DOC concentration did not vary significantly between B2 and B4 or between time periods. The time × treatment interaction was also not significant; thus, drainage water management did not influence DOC concentration (Table 5; Fig. 7). In contrast, DOC load was not different between B2 and B4, but it did vary significantly between time periods. The

time × treatment interaction for DOC load was significant, suggesting that drainage water management significantly influenced DOC loads at B4 (Table 5; Fig. 7). Comparing observed and predicted DOC loads for B4 showed that drainage water management decreased DOC load by 6.8 kg ha−1 (26%) between January 2009 and March 2011.

166

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

Table 4 Annual DOC load from individual subsurface tile drains and the watershed outlet. Year

Field tiles

Tile mains

County main

All tile drainage

Watershed outlet

B2

B4

B8

B5 (kg ha−1 )

B6

B3

2005 2006 2007 2008 2009 2010 2011a 2012

26.7 21.0 21.2 14.7 7.2 8.6 6.5 –

28.2 18.9 19.2 18.2 6.6c 6.7c 4.3c –

42.6 29.2 30.6 39.7 21.7 23.5 21.3 –

46.9 46.0 60.4 66.0 30.2 31.8 32.4 –

21.2 18.6 28.6 31.1 14.4 16.2 9.5 –

17.6 11.9 11.1 21.5 15.2 20.4 14.1 –

21.7 16.5 18.5 26.1 15.5 19.6 14.0 –

71.7 40.2 44.8 46.1 27.9 28.0 68.2 24.1

Avg.b

16.6

14.6

31.2

46.9

21.7

16.3

19.6

43.9

a b c

B1

Data for subsurface tile drains only measured through mid-March. Average only includes years 2005 through 2010 for tile drains. Drainage water management implemented at B4 from 2009 through 2011.

Table 5 Repeated measures analysis of subsurface tile drain discharge, DOC concentration, and DOC load from B2 and B4 using generalized linear model techniques. Bolded p-values are significant at p < 0.05. A significant treatment (control-impact) × time (before-after) interaction indicates that drainage water management had an effect on subsurface tile discharge, DOC concentration, or DOC load at B4. Effectsa

F-value

p-value

Tile discharge C vs. I B vs. A CI vs. BA

0.05 3.36 11.69

0.822 0.068 0.007

DOC concentration C vs. I B vs. A CI vs. BA

0.80 3.25 0.67

0.372 0.072 0.412

DOC load C vs. I B vs. A CI vs. BA

0.26 8.70 4.84

0.611 0.004 0.029

a Main effects: control vs. impact (C vs. I) and before vs. after (B vs. A); Interaction effect: control-impact vs. before-after (CI vs. BA).

Fig. 6. Ranked daily DOC load versus proportion of the study period (watershed outlet: January 2005-December 2012; subsurface tile drains: January 2005-March 2011).

4. Discussion 4.1. Effect of subsurface tile drains on watershed DOC export Quantifying the cumulative DOC load from all of the subsurface tile drains in an agricultural headwater watershed highlights the important role subsurface tile drainage plays in watershed DOC export and demonstrates the utility of monitoring hydrology and DOC loading at the headwater scale. In Watershed B, DOC load

from 6 subsurface tile drains comprised 33% of mean monthly DOC load from the watershed outlet, with cumulative DOC loads ranging from 15.5 to 26.1 kg ha−1 . Royer and David (2005) used DOC and discharge data from a field study by Kovacic et al. (2000) to estimate subsurface tile drain DOC loads in the Embarras River watershed in Illinois. They calculated that subsurface tile drains were contributing 9.4 kg ha−1 yr−1 (4–18 kg ha−1 yr−1 ) to the River, which was similar to the average DOC export from the watershed (9.5 kg ha−1 yr−1 ). Similarly, Dalzell et al. (2011) showed that areanormalized annual DOC export increased from 1.9 kg ha−1 (plot scale) to 7.5 kg ha−1 (Minnesota River basin) and suggested that subsurface tile drains have a larger influence on watershed DOC export at smaller scales compared to larger scales with additional sources of DOC. While numerous studies have quantified DOC loads from either subsurface tile drains or tile-drained watersheds, few have attempted to bridge the gap between scales. Prevailing processes controlling DOC at the field and watershed scales are not the same; yet, it is common practice to extrapolate field-scale results to the watershed scale. Conversely, it is difficult to infer processes occurring throughout a watershed based on a single monitoring point at the outlet. More studies focused on quantifying the relationship between field scale processes and watershed response are needed in the future. Patterns in DOC flux from both subsurface tile drains and the stream in Watershed B were largely controlled by hydrology, as >80% of DOC loading occurred during only 20% of the study period and seasonal DOC losses followed patterns in seasonal discharge. Previous work in this watershed has shown that under wet antecedent conditions (e.g., winter and spring) subsurface tile drain and watershed discharge responded rapidly to precipitation, which produced flashy hydrographs (King et al., 2014). While DOC concentration was not found to be strongly correlated to discharge in the current study, high flow events resulted in large DOC losses from subsurface tile drains to the stream channel and ultimately at the watershed outlet. In contrast, under dry antecedent conditions (e.g., summer) precipitation did not yield a similar hydrologic response, as tile drains often ceased to flow for periods of time and watershed response was greatly dampened (King et al., 2014). The intra-annual variability in subsurface tile drain and watershed hydrology not only resulted in seasonal differences in DOC loading, but also resulted in subsurface tile drains comprising a substantially larger portion of watershed DOC export during the wet periods (i.e., winter; 55%) compared to dry periods (i.e., summer; 29%). For the majority of the year, subsurface tile drain DOC concentration was often less than stream DOC concentration in Watershed B suggesting that stream DOC was being diluted by inputs from subsurface tile drains. The exception to this trend occurred during the summer and early-fall when subsurface tile drains tended

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

167

Fig. 7. Relationship between subsurface tile drains B2 (control) and B4 (treatment) showing the effect of drainage water management on (A) DOC concentration and (B) DOC load. From 2005–2008, both subsurface tile drains were free draining. Drainage water management was implemented at B4 from 2009 to 2011, while B2 remained in free drainage.

to function as a source of DOC to the stream (e.g., subsurface tile drain DOC concentration > stream DOC concentration). One possible explanation for the increased DOC concentrations in subsurface tile drains during these periods is the preferential transport of DOC through macropores from the soil surface to the drainage network (Vidon et al., 2008). Indeed, DOC concentration tended to be greater at subsurface tile drains B3 and B5 compared to other subsurface tile drains. Both of these subsurface tile drains had surface inlets connected to the drainage network which illustrates the potential influence of bypass flow on subsurface DOC concentration. The large variations in stream DOC concentration (0.1–44.4 mg L−1 ) and the general dilution of DOC concentration due to subsurface tile drain inputs observed in Watershed B were similar to observations reported elsewhere (e.g., Royer and David, 2005; Ruark et al., 2009), but mean DOC concentrations were greater in Watershed B compared to values in previous studies. Mean DOC concentrations from tile-drained fields and watersheds have generally been found to range from 1.5 to 5.0 mg L−1 (Ruark et al., 2009; Warrner et al., 2009; Dalzell et al., 2011), whereas mean DOC concentrations in Watershed B were between 5.7 and 8.2 mg L−1 . To elucidate the specific factors influencing DOC concentration and differences between

the current study and previous research, detailed information on agricultural management practices including tillage practices, crop type, crop yield, and nutrient management, as well as changes in soil carbon over the study period are required. Indeed, differences in vegetation and soil carbon to nitrogen ratio (Aitkenhead and McDowell, 2000), cultivation (Lal, 2004), and drainage intensity (Dalzell et al., 2011) have been suggested to influence DOC concentration. While information on nutrient management (i.e., crop type, tillage, etc.) was available for some fields during portions of the 8 year monitoring period (see Table 1), it was difficult to obtain management records from private landowners at the watershed scale over the course of a long-term research project (King et al., 2016). Based on the data in the current study, it appears that DOC export from tile-drained fields was linked to hydrology and drainage design (i.e., presence of surface inlets). The influence of other management factors on DOC concentration and load in subsurface drained fields and watersheds is a critical research need. The ability to obtain detailed management information from fields should therefore be considered a priority in watershed water quality studies. Results from the current study clearly show the significance of subsurface tile drains on watershed DOC export, but it is also

168

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

important to note the influence of other uncharacterized sources of DOC on watershed DOC budgets. On average, cumulative subsurface tile drain DOC loads accounted for one-third of the DOC flux from Watershed B, indicating that other sources such as groundwater seepage into the stream channel or in-stream production of DOC comprise the remaining two-thirds of watershed DOC export. The stream in Watershed B had dense growths of filamentous algae along the length of the channel, especially during the summer, which represents a potentially large source of DOC. Filamentous algae in agricultural streams can exceed 200 g m−2 dry mass (Schaller et al., 2004), with in-stream production exceeding 15 g C m−2 d−1 in open-canopied systems during the summer (Wiley et al., 1990). In addition, stream flow in Watershed B continued even when subsurface tile drains were not flowing, suggesting that groundwater seepage into the channel was occurring. Vidon et al. (2008) showed that groundwater was an important source of DOC to agricultural streams during baseflow. 4.2. Managing DOC loss from tile-drained fields Managing carbon has not traditionally been a priority in Midwestern agricultural landscapes, as management practices have typically focused on decreasing nitrogen, phosphorus, and sediment losses (e.g., Stanley et al., 2012). However, increasing promotion of soil health and carbon sequestration initiatives has raised awareness of carbon management. Results from the current study showed that implementing drainage water management decreased DOC losses from a tile-drained field by 26% (6.8 kg ha−1 ) over a 27 month period compared to an adjacent free draining subsurface tile outlet. This finding is consistent with results from studies examining the effectiveness of surface drain blocking in peatlands to limit DOC inputs to streams (Wallage et al., 2006; Armstrong et al., 2010). For example, Armstrong et al. (2010) found in a national survey of 32 surface drains in peatlands across the U.K. that DOC in blocked surface drains was, on average, 28% less than from unblocked surface drains. Thus, results from the current study demonstrate that drainage water management can be an effective practice for reducing DOC loss from tile drains. Reductions in DOC load from subsurface tile drain B4 were primarily due to decreases in discharge, as drainage water management did not significantly affect DOC concentration. The 22% decrease in subsurface tile discharge resulting from implementation of drainage water management was similar to that observed in previous studies. Adeuya et al. (2012) observed a 15–24% reduction in subsurface tile discharge from a field with drainage water management compared to a field with free drainage. Gunn et al. (2015) also measured a 40–100% decrease in subsurface tile drain discharge from six paired fields in northwestern Ohio. Differences in discharge reductions among studies can be attributed to the height of the subsurface tile outlet. In the current study, the height of the subsurface tile outlet was raised to 45 cm from the soil surface and only lowered to accommodate spring and fall field operations. A review by Skaggs et al. (2012) suggests that the outlet height is typically raised to a height of 30 cm in drainage water management studies, as was done in the study by Gunn et al. (2015). The closer the outlet height is set to the soil surface, the greater potential to decrease subsurface tile discharge; therefore, increasing the outlet height in the current study may have resulted in larger reductions in both subsurface tile drainage and DOC load. The U.S. Midwest is expected to experience greater intensity rainfall events along with an increase in spring rainfall during the mid- and late 21st century under projected future climate scenarios (e.g., Pryor et al., 2014). Increases in spring rainfall and greater intensity rainfall has the potential to result in increased DOC export from tile-drained watersheds. Drainage water management is one strategy to mitigate the impact of these projected changes on DOC

loss and ensure agricultural resiliency to future climate. Pease et al. (2017) employed the DRAINMOD hydrologic model to simulate subsurface tile drain discharge from fields in Ohio under future climate patterns. The authors found that the performance and benefits of drainage water management as an agricultural practice to control subsurface discharge were still evident under the projected climate change of the next century. Based on findings from the current study, drainage water management should therefore support DOC load reductions from tile-drained fields over the long-term. Results from the current study and others (e.g., Royer and David, 2005; Vidon et al., 2008; Warrner et al., 2009; Dalzell et al., 2011) have shown that substantial inputs of carbon in agricultural watersheds are from terrestrial sources. Thus, management efforts must consider carbon sources and delivery flow pathways in order to be effective, which can be challenging in watersheds with subsurface tile drainage where hydrologic flow paths have been altered and promote short-circuiting. Drainage water management may provide an opportunity to reconnect upland agricultural fields with riparian zones. Reconnecting fields and riparian zones has the potential to influence stream DOC dynamics via a number of direct and indirect methods including altered DOC inputs, insulation, and increased water residence time (e.g., Stanley et al., 2012). While DOC transport in groundwater was not measured in the current study, it is likely that drainage water management promoted lateral seepage of water through the riparian zone, as well as altered the quality and quantity of DOC delivered to the stream channel. Additional assessment of drainage water management and other practices, such as saturated buffers (e.g., Jaynes and Isenhart, 2014) that reconnect subsurface tile drainage to riparian zones, is needed to better understand DOC dynamics resulting from implementation of these practices.

5. Conclusions Quantifying the contributions of subsurface tile drainage to watershed DOC export is an important step in assessing the impact of agricultural management practices on stream water quality in artificially drained landscapes. Findings of the current study indicate that between 1 and 82% (mean = 33%) of monthly watershed DOC export was comprised of DOC from the subsurface drainage network in Watershed B. Much of the DOC loading occurred during high flow events, with greater contributions of DOC from subsurface tile drains occurring during the winter and spring seasons compared to the summer. Results of this study also demonstrate the utility of maintaining a long-term monitoring program in agricultural headwater watersheds to understand seasonal and annual variations in watershed hydrology and DOC loading, and confirm previous work in Midwestern agricultural watersheds that have suggested that subsurface tile drains are a large source of DOC to streams in these landscapes. Implementing drainage water management at one of the subsurface tile drains in Watershed B showed that DOC losses to streams can be decreased through use of this practice compared to an adjacent free draining control site. Decreases in DOC from subsurface tile drains have important implications for both in-field and in-stream biogeochemical processing. Obtaining detailed soil and management information from tile-drained fields and watersheds will be critical to further understand the factors controlling DOC export in these landscapes. Future research on potential factors influencing DOC concentration and loading in tile-drained landscapes is therefore needed, as well as the quantification of other sources of DOC such as groundwater seepage and in-stream DOC production on DOC export from agricultural watersheds.

M.R. Williams et al. / Agricultural Water Management 192 (2017) 159–169

Disclaimer 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. Acknowledgments The authors would like to thank the land owners of the study sites who provided access to the fields; Larry Ufferman, Ed Miller, and Bret Bacon of the Delaware County (Ohio) Soil and Water Conservation District for helping to identify cooperators and collecting management data; Eric Fisher for his analytical expertise; and Ginny Roberts, Jeff Risley, Sarah Hess, Ann Kemble, and Phil Levison for their help in data collection and site maintenance. The authors would also like to acknowledge the continued support of The Ohio State University in hosting the USDA-ARS Soil Drainage Research Unit and providing academic access to faculty resources, libraries, and students. References Adeuya, R., Utt, N., Frankenberger, J., Bowling, L., Kladiviko, E., Brouder, S., Carter, B., 2012. Impacts of drainage water management on subsurface drain flow, nitrate concentration, and nitrate loads in Indiana. J. Soil Water Conserv. 67, 474–484. Aitkenhead, J.A., McDowell, W.H., 2000. Soil C:N ratio as a predictor of annual riverine DOC flux at local and global scales. Glob. Biogeochem. Cycles 14, 127–138. Armstrong, A., Holden, J., Kay, P., Francis, B., Foulger, M., Gledhill, S., McDonald, A.C., Walker, A., 2010. The impact of peatland drain-blocking on dissolved organic carbon loss and discolouration of water; results from a national survey. J. Hydrol. 381, 112–120. Dalzell, B.J., Filley, T.R., Harbor, J.M., 2007. The role of hydrology in annual organic carbon loads and terrestrial organic matter export from a Midwestern agricultural watershed. Geochim. Cosmochim. Acta 71, 1448–1462. Dalzell, B.J., King, J.Y., Mulla, D.J., Finlay, J.C., Sands, G.R., 2011. Influence of subsurface drainage on quantity and quality of dissolved organic matter export from agricultural landscapes. J. Geophys. Res. 116, G02023, http://dx.doi.org/ 10.1029/2010JG001540. Gunn, K.M., Fausey, N.R., Shang, Y., Shedekar, V.S., Ghane, E., Wahl, M.D., Brown, L.C., 2015. Subsurface drainage volume reduction with drainage water management case studies in Ohio, USA. Agric. Water Manage. 149, 131–142. Hinton, M.J., Schiff, S.L., English, M.C., 1997. The significance of storms for the concentration and export of dissolved organic carbon from two Precambrian Shield catchments. Biogeochemistry 36, 67–88. Jaynes, D.B., Isenhart, T.M., 2014. Reconnecting tile drainage to riparian buffer hydrology for enhanced nitrate removal. J. Environ. Qual. 43, 631–638. Kennedy, C.D., Bataille, C., Liu, Z., Ale, S., VanDeVelde, J., Roswell, C.R., Bowling, L.C., Bowen, G.J., 2012. Dynamics of nitrate and chloride during storm events in agricultural catchments with different subsurface drainage intensity (Indiana, USA). J. Hydrol. 466, 1–10. 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. King, K.W., Williams, M.R., Fausey, N.R., 2015. Contribution of systematic tile drainage to watershed-scale phosphorus transport. J. Environ. Qual. 44, 486–494. King, K.W., Williams, M.R., Fausey, N.R., 2016. Effect of crop type and season on nutrient leaching to tile drainage under a corn-soybean rotation. J. Soil Water Conserv. 71, 56–68. Kladivko, E.J., Brown, L.C., Baker, J.L., 2001. Pesticide transport to subsurface tile drains in humid regions of North America. Crit. Rev. Environ. Sci. Technol. 31, 1–62. Kovacic, D.A., David, M.B., Gentry, L.E., Starks, K.M., Cooke, R.A., 2000. Effectiveness of constructed wetlands in reducing nitrogen and phosphorus export from agricultural tile drainage. J. Environ. Qual. 29, 1262–1274. Lal, R., 2004. Soil carbon sequestration impacts on global climate change and food security. Science 304, 1623–1627.

169

Ledesma, J.L.J., Köhler, S.J., Futter, M.N., 2012. Long-term dynamics of dissolved organic carbon: implications for drinking water supply. Sci. Total Environ. 432, 1–11. 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 Manage. 92, 171–182. Mausbach, M.J., Dedrick, A.R., 2004. The length we go. J. Soil Water Conserv. 59, 97A–103A. Mulholland, P.J., 1997. Dissolved organic matter concentration and flux in streams. J. N. Am. Benthol. Soc. 16, 131–141. NCDC (National Climatic Data Center), 2016. http://www.ncdc.noaa.gov/cdo-web/ (Accessed 15 May 2016). NRCS (National Resource Conservation Service), 2003. Managing Soil Organic Matter: The Key to Air and Water Quality. Soil Quality Technical Note No. 5. US Department of Agricultural Natural Resource Conservation Service, Washington DC (Accessed 1 June 2016) http://www.nrcs.usda.gov/Internet/ FSE DOCUMENTS/nrcs142p2 050965.pdf. Ohio EPA (Ohio Environmental Protection Agency), 2004. Total Maximum Daily Loads for the Big Walnut Creek Watershed. Draft Report for Public Comment, Dec. 8, 2004. Division of Surface Water. State of Ohio Environmental Protection Agency, Columbus, OH. Pease, L.A., Fausey, N.R., Martin, J.F., Brown, L.C., 2017. Projected climate change effects on subsurface drainage and the performance of controlled drainage in the Western Lake Erie Basin. J. Soil Water Conserv. 72, 240–250. Peterson, B.J., Wolheim, W.M., Mulholland, P.J., Webster, J.R., Meyer, J.L., Tank, J.L., Marti, E., Bowden, W.B., Valett, H.M., Hershey, A.E., McDowell, W.H., Dodds, W.K., Hamilton, S.K., Gregory, S., Morrall, D.D., 2001. Control of nitrogen export from watersheds by headwater streams. Science 292, 86–90. Pryor, S.C., Scavia, D., Downer, C., Gaden, M., Iverson, L., Nordstrom, R., Patz, J., Roberson, G.P., 2014. Chapter18: midwest. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. US Global Change Research Program, Washington, DC, pp. 419–440. R Development Core Team, 2011. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Royer, T.V., David, M.B., 2005. Export of dissolved organic carbon from agricultural streams in Illinois, USA. Aquat. Sci. 67, 465–471. Ruark, M.D., Brouder, S.M., Turco, R.F., 2009. Dissolved organic carbon losses from tile drained agroecosystems. J. Environ. Qual. 38, 1205–1215. Schaller, J.L., Royer, T.V., David, M.B., Tank, J.L., 2004. Denitrification associated with plants and sediments in an agricultural stream. J. N. Am. Benthol. Soc. 23, 667–676. Skaggs, R.W., Fausey, N.R., Evans, R.O., 2012. Drainage water management. J. Soil Water Conserv. 67, 167A–172A. Smith, E.P., 2002. BACI design. In: Encyclopedia of Environmetrics. John Wiley & Sons, Ltd., Chichester (pp. 141–148). Stanley, E.H., Powers, S.M., Lottig, N.R., Buffam, I., Crawford, J.T., 2012. Contemporary changes in dissolved organic carbon (DOC) in human-dominated rivers: is there a role for DOC management. Freshw. Biol. 57, 26–42. Stone, W.W., Wilson, J.T., 2006. Preferential flow estimates to an agricultural tile drain with implications for glyphosate transport. J. Environ. Qual. 35, 1825–1835. Tomer, M.D., Meek, D.W., Jaynes, D.B., Hatfield, J.L., 2003. Evaluation of nitrate nitrogen fluxes from a tile-drained watershed in central Iowa. J. Environ. Qual. 32, 642–653. USDA SSURGO (Soil Survey Geographic Database), 2017. Natural Resources Conservation Service. United States Department of Agriculture, Available online at http://sdmdataaccess.sc.egov.usda.gov (Accessed 15 May 2017). Vidon, P., Wagner, L.E., Soyeux, E., 2008. Changes in the character of DOC in streams during storms in two Midwestern watersheds with contrasting land uses. Biogeochemistry 88, 257–270. Wallage, Z.E., Holden, J., McDonald, A.T., 2006. Drain blocking: an effective treatment for reducing dissolved organic carbon loss and water discolouration in drained peatland. Sci. Total Environ. 367, 811–821. Warrner, T.J., Royer, T.V., Tank, J.L., Griffiths, N.A., Rosi-Marshall, E.J., Whiles, M.R., 2009. Dissolved organic carbon in streams from artificially drained and intensively farmed watersheds in Indiana, USA. Biogeochemistry 95, 295–307. Wiley, M.L., Osborne, L.L., Larimore, R.W., 1990. Longitudinal structure of an agricultural prairie river system and its relationship to current stream ecosystem theory. Can. J. Fish. Aquat. Sci. 47, 373–384. Williams, M.R., King, K.W., Fausey, N.R., 2015. Contribution of tile drains to basin discharge and nitrogen export in a headwater agricultural watershed. Agric. Water Manage. 158, 42–50.