Development and evaluation of SWATDRAIN, a new model to simulate the hydrology of agricultural tile drained watersheds

It is important for watershed models to realistically simulate tile drainage flow and water table dynamics. Therefore, a new model, SWATDRAIN, was developed in this study by incorporating the DRAINMOD model into the Soil and Water Assessment Tool (SWAT) to better simulate surface and subsurface flow in tile-drained watersheds and also to improve the prediction of water table depth. This was accomplished by fully integrating the DRAINMOD model, which has been tested and widely used to simulate the performance of drainage and water table control systems on a continuous basis at field scale, into the SWAT model. The SWATDRAIN model was evaluated for a fully tile-drained watershed in eastern Ontario, Canada. The measured tile drainage outflow and water table depth data for the Green Belt watershed were used to evaluate the capability of the new model to simulate water balance for this fully tile drained agricultural watershed. Together with hydrographs, the Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS) and coefficient of determination (R2) statistics were used in evaluating the accuracy of SWATDRAIN to predict tile flow and water table depth in light of the measured values. Simulations were carried out over the period of 1991 to 1993; 1991 and 1992 data served as model calibration and 1993 data were used to validate the process. Model accuracy statistics for the monthly and daily water table depth over the validation period were, respectively, 0.86 and 0.70 for R2, 0.11 and 2.90 for PBIAS, and 0.80 and 0.67 for the NSE. Model accuracy statistics for events, monthly and daily tile drainage over the validation period were, respectively, 0.86, 0.88 and 0.70 for R2, 11.7, 17.26 and 23.85 for PBIAS, and 0.84, 0.86 and 0.62 for the NSE. The SWATDRAIN model was also applied to a partially tile-drained watershed in southern Ontario. Simulations were carried out from 1975 to 1983; data from 1975 to 1978 were used for model calibration and data from 1980 to 1983 were used for validation. The new model was able to adequately simulate the hydrologic response at the outlet of the watershed. Comparing the observed monthly and daily tile drainage with the model's output over the validation period returned R2 values of 0.75 and 0.62, PBIAS of 13.96 and 17.99 and modeling efficiency of 0.71 and 0.62. In this study, the effects of a drainage water management operational strategy on hydrology were simulated using SWATDRAIN in the Green Belt watershed in Ontario. The effects of drainage water management on subsurface drainage and surface runoff were predicted for a period of four years from 2004 to 2007. Implementing the controlled drainage strategy from June 15 to August 15 during the cropping season and also from November 1 to May 1 in the non-growing season resulted in a reduction of the average annual drain flow by 18%, while it increased the surface runoff in the order of 30%. The results showed that the surface runoff increase mostly happened during the snowmelt period in April and also it was slightly increased during the month of November. However, higher amount of surface runoff in flat watersheds during the snowmelt period may not cause a serious problem.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.