Final Report for GNC06-066
Drainage water management can reduce nitrate loss from drained fields while maintaining crop production.
This project measured the impact of drainage water management on drain flow and water table depth on three Indiana farms using the paired watershed approach.
Drainage water management was found to reduce drainage outflow during both summer and winter treatment, and maintain water table depth at higher levels than the conventionally-drained field. The impact of drainage water management on flow and water table depth were greater during the winter treatment period compared to the summer period.
Agricultural drainage plays a vital role in the cropping systems in the Midwest; however, conventionally-managed drainage has been found to increase nitrate loss through enhanced leaching of the soil profile(Tomer et al., 2003). Subsurface drainage is used to improve soil aeration, lower surface runoff through increased infiltration, and therefore improve crop growth and yield, but has been reported as a major source of nitrate loading to surface water in the Mississippi River Basin (Burkhart and James, 1999; Randall and Goss, 2001).
A number of practices are being studied that have potential to reduce nitrate losses from subsurface tile drains. Some of these include reduced fertilizer applications, cover and perennial crops, biofilters, and wetlands (e.g.`, Dinnes et al., 2002; Randall and Mulla, 2001). Drainage water management (also known as controlled drainage) is one of the practices that show great promise for Midwest flat land, because it has the potential to reduce nitrate loss from drained fields while maintaining drainage intensity critical to crop growth. It works by raising the water table level at which drainage occurs by raising the effective height of the drain outlet, which reduces drainage and consequently nitrate loss. A few studies in the Midwest (Drury et al., 1996; Fausey et al., 2004) have shown the potential for reducing nitrate loss under these conditions. On-farm studies are needed to test the practice under a greater variety of conditions, and better understand how the practice impacts the farming operation.
The field-scale research we have conducted with funds from the SARE grant was part of a larger project evaluating the impacts of drainage water management on corn and soybean cropping systems, yield, economic sustainability, soil quality and nitrate losses at the watershed scale. The larger project includes evaluating drainage water management potential impacts on soil physical properties, earthworms, plant growth, plant nitrogen content, yield and profitability. The focus of the SARE-funded project was on assessing the impact of this practice on water table and drain flow, through paired field on-farm trials of conventional and managed drainage on three Indiana fields.
The main objective of this project was to assess the impact of drainage water management on reducing tile outflow from drained fields, increasing available water and crop yield through paired-field studies on three farms in Indiana.
We also formulated intermediate and long-term desired outcomes, although they were not expected to be accomplished during the course of the project. The desired intermediate-term outcomes will:
1. Enable farmers to make more informed decisions on whether or not the adoption of the practice would be profitable for them, and
2. Enable federal, state, local, and non-governmental soil conservation and water quality organizations to make decisions about supporting drainage water management as a best management practice.
The desired long-term outcome is that farmers will adopt the practice (if it is determined to be successful at increasing yield, promoting improved water quality and reducing nitrate losses).
We worked with three Indiana farmers on this on-farm study. One had already become interested in managing drainage water through conversations with a dynamic drainage contractor, based on the potential yield impacts. The other two were identified for this research/demonstration project with the help of Extension staff and other partners. Each farmer agreed to the installation of water control structures, and to manage the entire project area with the same corn hybrid or soybean variety and similar pest and fertilizer management during the entire project period. One subfield on each site was used as control and the other as treatment, meaning that one subfield was subjected to drainage water management while the other was allowed to flow without constraint. The treatment and control was randomly assigned to subfield within each paired study. Both subfields had similar soils, drainage systems, and management histories following the split-field research design. The farms are referred to as Sites 1, 2, and 3. Sites 1 and 2 are in White County, and Site 3 is in Montgomery County.
Drain flow measurement
Flow was measured at these locations using three different methods.
(1) The modified circular flume(Cooke et al., 2004) had smaller vertical pipes within the larger drainage pipe that created critical flow. Pressure transducers were used to measure the water level in flumes. At one location there was an additional smaller vertical pipe used to determine downstream depth with the use of an additional pressure transducer following the method of Cooke et al. (2004). Although laboratory testing results were good, field measurements indicated the need for the development of more site-specific rating curves, that could vary depending on outlet submergence.
(2) The insertion electromagnetic flow meters (SeaMetrics Ex100; www.seametrics.com)were installed at two sites in a U-shaped section in the drainage pipe to create continuous full pipe flow conditions, for which there was a constant flow area and velocity measured to determine flow. The insertion flow meters had a minimum flow of 31 gallons per minute, but were very effective at measuring high flow rates.
(3) The area-velocity meters (Flo-Tote 3; www.marsh-mcbirney.com) also used an electromagnetic flow meter, together with a level sensor to measure water level in the pipe, and were the most effective at measuring flow at both low and high drain flow rates. The use of three flow-measurement methods was due to problems encountered with the first method, and also allowed for more complete data with flow determined by two different methods at each site.
Water table measurement
Observation wells were installed at all sites to compare the long-term water table depth under free drainage and drainage water management. Wells were placed at the midpoint between two tile lines to avoid the drawdown near tiles, and more than two times the tile spacing away from the structure, to minimize edge of field effects. Wells were drilled by hand to a depth of approximately 2 meters in the cropped portion of fields. Each well was equipped with a water level logger (Global water, WL16) that measured the water table every hour. The water table sensors were removed for field operations (planting and harvesting) twice per year, limiting available water table data.
The research was designed with the intention of applying the paired watershed method as described in Clausen and Spooner (1993). Both the treatment (drainage water management) and control (free drainage) at each site was allowed to flow without restriction for a period of time known as the calibration period. During this period the natural process of the drainage and water table depth were observed for each subfield. Data from the calibration time period were used to establish baseline relationships between the two subfields, needed for determining the treatment effects. The paired watershed approach requires the development of a linear model that relates the water table depth (or drain flow) for the controlled subfield to the corresponding water table depth (or drain flow) for the freely drained subfield. The treatment effect is determined by significant differences in the slope and or intercept for the calibration and treatment periods.
A check for normality was conducted through the visual analysis of the normal probability plot (QQ plot) as a means of verifying the assumption. Both water table depth and drain flow was found to approach normality for all research sites.
Drain flow and water table were found to be autocorrelated using the proc autoreg procedure in SAS, therefore violating the ANOVA assumptions of normally distributed residuals, equality of variances and independence. The repeated measures mixed method (Littel et al., 1998) was therefore used to conduct the paired field analysis. The repeated measures mixed method accounts for the correlation between data points and can therefore be used to analyze measurements that are repeated over time or space(Cheng et al., 2005). The repeated measured mixed method requires special attention be paid to the sequential nature of the data being measured. Treatment effects were determined by utilizing the spatial power function to model the covariance structure.
In addition to testing the differences in the intercepts and slopes of the calibration versus treatment regressions, the effect of drainage water management was also determined by percent change in mean water table depth and drain flow, and visually using bivariate plots of paired observations.
The results showed reduced drain outflow and water table depths closer to the soil surface for the managed subfield compared to the free drainage subfield. Water table depth and drain flow were measured from 2005-2008 and were adequate for conducting paired analysis.
Water table depth /flow comparisons
For Site 1, observed drain flow data were available for 3/6/2007 to 12/31/2008. The free drainage subfield had higher peak flows compared to the free drainage field. During winter, spring and fall of 2008 the managed subfield had low constant flow while the free drainage subfield had higher flow. The managed subfield water table started out closer to the soil surface during the summer and then dropped below that of the free water table midway through the growing season. The water table was expected to fall during the summer for the managed field as well as the unmanaged field because of evaporation, vertical seepage and crop usage. However the lowering of the managed water table below that of the free drainage field was unexpected because of the assumed similarity in the hydrology of subfields. This observation indicates that there was more drainage, possibly deep seepage from the managed field compared to free drainage field. During the winter the water table at the managed subfield appears to be higher than that of the free subfield for most of the time.
Observed drain flow data for Site 2 were available between 6/30/2006 to 12/31/2008. Both drains at Site 2 had similar flow except during the winter when the managed drain appeared to have lower flow levels than the free drainage subfield. The water table depth for Site 2, available from 9/20/2005 to 12/31/2008, showed a consistently lower water table for the winter of 2005-2006 for the managed subfield, similar levels among fields 2006-2007 and consistently higher level in the managed field for 2007-2008. During summer the water table in free subfield peaked at lower depths below the surface compared to the managed subfield and there appeared to be very little difference in the water table recession. The managed field was expected to have higher water table peaks when the field was being controlled due to drainage reduction. The managed subfield tends to have water table that was closer to the surface when controlled in the winter.
Site 3 had the longest flow record of all research sites (01/08/2005 to 12/31/2008) because data collection was initiated earlier there and the site had less equipment failures. During both managed and unmanaged periods the free drainage subfield of Site 3 had higher peak flows than the managed subfield. This was attributed to the constant submergence of the managed tile during high flow periods compared to the lack of submergence at the free drainage outlet. Observed water table depths for Site 3 from 7/05/2005 to 12/31/2008 showed the managed field water table to recede slower than the free drainage during the summer controlled period of 2006. The water table level in managed subfield was maintained at depth closer to the surface than free drainage subfield for most of the summer and winter of 2006 through 2008 when the water table was being managed.
Flow and water table depth were found to be autocorrelated to a maximum of 46 and 41 days respectively. The spatial correlation estimate ranged from 0.47 to 0.87 for flow and 0.88 to 0.99 for water table depth indicating a very strong correlation in the measured data and supports the use of the spatial power function to model the covariance structure. The repeated measured mixed model allowed the determination of the effect of drainage water management while accounting for the autocorrelation in water table depth and flow.
The slopes and/or intercept of the flow regressions were significantly different (p<0.05) for the winter treatments for all research sites. For summer the intercept of flow for Site 2 was significantly different (p<0.1), Site 1 did not have significant differences in slope or intercept for the summer period, while Site 3 flows had significant differences (p<0.05)in the intercept and slope. For both the winter and summer treatments there were significant differences (p<0.05) in either the slope and or intercept of water table treatments for all sites except for Site 3, which was significant at a p-value of 0.1.
Drainage water management was effective in reducing drainage outflow by 23% at Site 1, 15% at Site 2 and 25.8% at Site 3 for the summer treatment periods. During the winter period, drainage water management reduced outflow for Site 1 and 2 by 49.7% and 24%, respectively, while increasing outflow by 2% at Site 3. . Drainage water management was successful at maintaining water table depths closer to the soil surface compared to the free drainage field. During the winter there was a 71.7%, 36.5% and 12.3% rise in water table depth for Site 1, 2 and 3, respectively. For summer treatment Site 1 and 2 water table depths were lower compared to the free drainage subfield by 37.5% and 4.3%, respectively. Only Site 3 water table depth showed a positive effect of the practice during the summer treatment with 8% rise in water table depth.
Analysis of the effects of drainage water management on both flow and water table depth from agricultural fields showed that there was a greater effect of the practice in the winter treatment period than the summer. Although Site 3 was the only site that had significant differences in the slope of the regression for summer treatment, the average change in flow was substantial, ranging from 0.15 mm/day to 0.39 mm/day. For the winter treatment the effect ranged from -0.05 mm/day to 0.81 mm/day. The average change in water table depth ranged from 0.08 m to 0.33 m with the greatest differences occurring during the winter treatment period. Managing the flow during the summer did not have the desired effect on water table depth; however, analysis of water table depth averages the water table depth for the entire growing season. The greatest influence of the practice occurs in the earlier part of the season when the water table was above the drainage tile.
Burkhart, M. R. and James, D. E. (1999). Agricultural-nitrogen contributions to hypoxia in the Gulf of Mexico. Journal of Environmental Quality 28: 850-859.
Cheng, J., Olbricht, G., Gunaratna, N., Kendall, R., Lipka, A., Paul, S. and Tyner, B. (2005). Purdue SCS reference manual on mixed models. Purdue Statistical Consulting Service: Purdue University, W. Lafayette, IN.
Clausen, J. C. and Spooner, J. (1993).Paired watershed study design. USEPA, Vol. 841-F-93-009. Washington, DC: United States Environmental Protection Agency.
Cooke, R. A., Wildman, T. A. and Northcott, W. J. (2004).Accurate, low-cost instrument for measuring subsurface drain flow under submerged outlet conditions. Proceedings of the Eighth International Drainage Symposium, 332-339 (Ed R. Cooke). Sacramento, California, USA: ASAE 701P0304.
Dinnes, D. L., Karlen, D. B., Jaynes, D. B., Kaspar, J. L., Hatfield, J. L., Colvin, T. S. and Cambardella, C. A. (2002). Nitrogen management strategies to reduce nitrate leaching in tile-drained Midwestern soils. Agronomy Journal 94: 153-171.
Drury, C. F., Tan, C. S., Gaynor, J. D., Oloya, T. O. and Welacky, T. W. (1996). Influence of controlled drainage-subirrigation on surface and tile drainage nitrate loss. Journal of Environmental Quality 25: 317-324.
Fausey, N. R., King, K. W., Baker, B. J. and Cooper, R. L. (2004).Controlled drainage performance on Hotyville soil in Ohio. Drainage proceedings of the eighth Intenational Symposium, 84-88 (Ed R. A. Cooke). ASAE Publication No. 701P0304.
Littel, R. C., Henry, P. R. and Ammerman, C. B. (1998). Statistical analysis of repeated measures data using SAS proceedure. Journal of Animal Science 76(4): 1216-1231.
Randall, G. W. and Goss, M. J. (2001).Nitrate losses to surface water through subsurface, tile drainage. In R.F. Follett and J.L. Hatfield (Eds). Nitrogen in the Environment: Sources, Problems, and Management., 95-122.
Randall, G. W. and Mulla, D. J. (2001). Nitrate nitrogen in surface waters as influenced by climatic conditios and agicultural practices. Journal of Environmental Quality 30: 337-334.
Tomer, M. D., Meek, D. W., Jaynes, D. B. and Hatfield, J. L. (2003). Evaluation of nitrate fluxes from tile-drained watershed in Central Iowa. Journal of Environmental Quality 32: 642-643.
Educational & Outreach Activities
Field days and tours:
• On June 21, 2007 the Conservation Information Technology Center (CTIC) board of directors visited Site 3 as part of their 2007 Conservation Tour. The cooperating farmer discussed his experience and opinions about the practice with this influential group.
• A public Field Day was held at Site 1 on September 2, 2008 targeting the agricultural community. The cooperating farmers from both Sites 1 and 2 spoke about their experience with the practice, as well as representatives from Purdue, USDA-NRCS, and the Agricultural Drainage Management Coalition.
• In March 2008 the NCERA 207 Multi-state Committee “Drainage design and management practices to improve water quality” and the Agricultural Drainage Systems Task Force (ADMS-TF) meetings were held concurrently a Purdue University and included a field trip to Site 1 to demonstrate our drainage water management research design.
Adeuya, R., J. Frankenberger, E. Kladivko, L. Bowling. 2007. The effect of controlled drainage on water table depth at research sites in Indiana. ASABE Annual International Meeting, June 17-20. Minneapolis, MN.
Frankenberger, J.R., Kladivko, E., Adeuya, R., Utt, N., Bowling, L. and Carter, B., 2008. Determining the hydrologic impacts of drainage water management in Indiana, USA. 10th International Drainage Workshop; Helsinki, Finland/ Tallinn, Estonia. July 6-11th 2008.
A doctoral dissertation and additional conference publications will results from this research. Publications in peer reviewed journals are also anticipated. This research will be used to prepare extension materials for field days and other outreach programs with the state of Indiana and beyond.
Drainage water management reduced the drainage outflow significantly during the high flow periods (winter). Because nitrate concentration is usually constant, the reduction in flow means that the nitrate loss was significantly reduced as well. This has positive effects on reducing nitrate losses, and shows the potential for reducing nitrate concentrations downstream. Other studies are investigating the potential of drainage water management for achieving nitrate reductions at the watershed scale and even at the scale of the hypoxic zone in the Gulf of Mexico, and the field-scale results of this and other studies in the region provide needed data to estimate these possible nitrate reductions.
Through field days and other methods, additional farmers within our research area are becoming more interested in adopting the practice. One additional private farm was added to our research since the project began. The cooperating farmers are generally pleased with the practice and actively participate in field activities, giving assistance when needed. In the future I expect that more farmers within the state of Indiana will adopt this practice as the findings of this research becomes available. The current farmers on the research project are expected to continue maintain and using the practice as part of their farm operations.
An economic analysis was not conducted as part of this research. However, drainage water management is expected to increase yield through the increase availability of water during the growing season. Site-specific crop yield data were collected by the farmers at each site each year using GPS-enabled combine yield monitors. Yield analysis as well as detailed economic analysis from theses research location will be published by other members of the larger project in the coming years.
The economics of the practice are also influenced by cost-share programs available to farmers for the installation and management of this practice through the USDA National Resource Conservation Service (NRCS) Environmental Quality Incentives Program (EQIP) and Conservation Security Program (CSP). These payments and incentives are needed to support drainage water management during the winter, when there is no yield or economic benefit to the farmer.
The farmers that were a part of this project are generally enthusiastic about the possibilities for increased crops yields and improving the quality of water leaving their fields. I expect that these farmers will continue to manage the drainage outflow from their fields when our research is completed. As more field days are held and the practice becomes more widely known, others are likely to adopt it.
Day-to-day maintenance of a drainage water management is fairly simple. The control is set at specific heights that are reflective of the desired depth below the surface of the lowest spot in the field, at the beginning of the growing season and during the winter. The controlled levels are changed for planting and harvesting to make the soil suitable for machinery. Only in events of severe storm events with high rainfall intensities do farmers need to check control levels to prevent flooding in fields.
Areas needing additional study
Although drainage water management was shown to reduce drainage outflow and maintain water table depth at higher levels compared to free drainage subfields, more needs to be done to comprehensively understand the water balance concerns with the use of the practice. Studies that are able to determine the field boundary conditions as it relates to edge of field water loss would provide a needed additional dimension. Additional water table depth measurements would help in analyzing the yield effects associated with drainage water management. A comprehensive nitrogen balance study that involved intensive monitoring (daily measurements) of nitrogen concentrations in outflow (drainage outflow and edge of field flow losses) coupled with soil and plant nitrogen measurements would aid researchers in the affirmation of the time when managing the drainage flow is most critical to reducing nitrate losses.