|Mulla, David - University Of Minnesota|
|Ale, Srinivasulu - Texas A&M Agrilife|
Submitted to: Journal of Environmental Quality
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/10/2013
Publication Date: 11/15/2013
Citation: Moriasi, D.N., Gowda, P., Arnold, J.G., Mulla, D.J., Ale, S., Steiner, J.L., Tomer, M.D. 2013. Evaluation of the Hooghoudt and Kirkham tile drain equations in SWAT to simulate tile flow and nitrate-nitrogen. Journal of Environmental Quality. 42:1699-1710.
Interpretive Summary: Subsurface tile drainage is a commonly used agricultural practice to enhance crop yield in poorly drained but highly productive soils in many regions of the world. However, the presence of subsurface tile drainage systems also expedites the transport of nitrate-nitrogen (NO3-N) and other chemicals to surface waters. Hydrologic and water quality models such as the Soil and Water Assessment Tool (SWAT) are widely used to simulate tile drainage systems at various spatial scales. However, the soil profile moisture method used by SWAT to compute surface runoff does not correctly partition surface runoff and tile flow for poorly drained soils and mildly-sloped agricultural areas. This has huge implications on the prediction of soil erosion and phosphorus losses while greatly affecting the partitioning of NO3-N losses between surface runoff and subsurface tile drainage. This also has major implications on time to peak flow and magnitude of peak flow, which are important in channel processes. In this study, modifications were made to correct problem in SWAT after which the Revised SWAT model was calibrated and validated using subsurface tile flow and NO3-N using long term monitoring data from southern Minnesota plots to ensure that water and nitrogen budgets were properly simulated. The method was revised to include a factor to account for variable tile drainage systems, mild slopes, and drainable properties of shallow poorly drained soils. Study results indicated that the modifications improved prediction of tile flow and NO3-N losses associated with tile drainage by 54% and 39%, respectively, while simulating minimal surface runoff as reported for the study area. The improved Revised SWAT model will lead to more accurate results of the impacts of water management practices to reduce contribution of NO3-N loadings to water bodies.
Technical Abstract: Subsurface tile drains in agricultural systems of Midwest U.S. are a major contributor of nitrate-N (NO3-N) loadings to hypoxic conditions in the Gulf of Mexico. Existing soil moisture retention parameter computation algorithm in the widely used Soil and Water Assessment Tool (SWAT) model is known to have errors in the partitioning of surface runoff and tile flow in poorly drained soils and mildly-sloped agricultural areas such as those in the Mid-western U.S. In this study, long-term (1983-1996) monitoring data from southern Minnesota was used to revise and evaluate the SWAT model for accurately estimating surface and subsurface tile drain flows and associated NO3-N losses. Specific objectives of this study include: 1) modifications made to the traditional soil moisture retention parameter calculation method in the currently distributed version of SWAT to account for the effects of tile drainage, and slope on the computation of surface runoff using the curve number (CN) method in poorly drained and mildly-sloped agricultural watersheds (referred herein as Revised SWAT) and 2) calibration and validation of the SWAT and Revised SWAT models for tile flow and associated NO3-N losses from poorly drained soils and mildly-sloped agricultural areas using long term monitoring data. Monitoring data were divided equally for model calibration and validation purposes. A retention parameter adjustment factor was incorporated to account for tile drainage and adjust for slope, which are not accounted for in tile drained and mildly-sloped agricultural plots, fields, or watersheds. Comparison of predicted monthly tile flow and associated NO3-N losses by currently distributed version of the SWAT and the Revised SWAT models with measured data indicated a better fit for predicted values using the Revised SWAT model when using the CN input values for the study area. For the calibration period, the Revised SWAT model simulated tile flow and NO3-N losses within 1% and 4% of the observed data, respectively. For the validation period, it simulated tile flow and NO3-N losses within 5% and 7%, respectively, of the observed values. Revised SWAT predicted water and NO3-N budgets were in the same order of magnitude of the measured or those reported in the literature.