|VAZQUEZ-AMABILE, GABRIEL - National University Of Laplata|
|ENGEL, BERNARD - Purdue University|
Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/7/2011
Publication Date: 8/1/2011
Citation: Moriasi, D.N., Arnold, J.G., Vazquez-Amabile, G.G., Engel, B.A. 2011. Shallow water table depth algorithm in SWAT: Recent developments. Transactions of the ASABE. 54(5):1705-1711.
Interpretive Summary: In a recent study, a new water table depth (wtd) prediction method called Modified DRAINMOD was developed and incorporated into the Soil and Water Assessment Tool (SWAT), a continuous-time physically-based watershed-scale hydrologic model, in order to improve the simulation of wtd dynamics. Simulating wtd dynamics accurately is essential because the proximity of wtd to the soil surface impacts drainage, crop development, agricultural chemical transport, soil salinity, and farm machine trafficability. In this approach drainage volume is converted into water table depth using a constant conversion factor that is determined for each field (or hydrologic unit HRU) in SWAT by a manual adjustment process that works to fit, as closely as possible, the simulated wtd fluctuation pattern to the measured wtd fluctuation pattern. However, at the watershed-scale where there are many fields (or HRUs), it is difficult if not impossible to determine an optimum water table conversion factor value for each HRU through the manual adjustment process. The objectives of this study were to: 1) revise the Modified DRAINMOD wtd approach in SWAT so that water table conversion factor is automatically computed by the model as a function of soil physical properties, in order to eliminate determination of this factor through the manual adjustment process; and 2) evaluate the revised Modified DRAINMOD model within SWAT using measured water table depth data from three observation wells located in forest fields without tile drainage within the Muscatatuck River basin in southeast Indiana. Based on these model performance results, there were no significant differences between the wtd simulated using the manually determined constant and the automatically computed variable water table conversion factor values. However, the wtd dynamics simulated using the automatically computed variable factors were on average 10% to 18% closer to the measured wtd dynamics than the wtd dynamics simulated using the manually determined constant water table factors. Automatically computed variable water table factors will enable this new more accurate shallow wtd algorithm in SWAT to be used at the watershed scale.
Technical Abstract: Knowledge of the shallow water table depth (wtd) is crucial in many studies including determination of optimum irrigation and drainage management systems for agricultural production, farm machine trafficability, and water quality due to agricultural chemical transport and soil salinity. Therefore, it is essential for hydrologic models to accurately simulate wtd. Recently, a new shallow wtd algorithm that relates drainage volume (vol) to wtd was incorporated into the Soil and Water Assessment Tool (SWAT) model. Water table depth is computed as a function of vol and a water table factor (wt_fctr), which converts vol into wtd. The constant wt_fctr is currently a calibration parameter. However, at the watershed-scale where there are many fields (hydrologic response units, HRUs), it is difficult if not impossible to determine an optimum wt_fctr value for each HRU through the calibration process. The objectives of this study were to: 1) modify the new shallow wtd algorithm in SWAT so that wt_fctr is automatically computed by the model as a function of soil physical properties, in order to eliminate determination of wt_fctr through the calibration process; and 2) evaluate the modified wtd algorithm within SWAT using measured water table depth data from three observation wells located in forest fields without tile drainage within the Muscatatuck River basin in southeast Indiana. On average the calibrated wt_fctr yielded daily calibration and validation Nash-Sutcliffe efficiency (NSE) values of 0.64 and 0.41, respectively, the percent bias (PBIAS) values of -13% and -4%, respectively, and root mean square error (RMSE) values of 0.41 m and 0.59 m, respectively, while the automatically computed variable wt_fctr yielded NSE values of 0.66 and 0.58, respectively, PBIAS values of 4% and 10%, respectively, and RMSE values of 0.40 m and 0.50 m, respectively, for the three observation wells. Based on these model performance results, there were no significant differences between the wtd simulated using the manually calibrated constant and the automatically computed variable wt_fctr values. Automatically computed variable wt_fctr will enable this new shallow wtd algorithm to be used at the watershed scale.