Application of HYDRUS (2D/3D) for Predicting the Influence of Subsurface Drainage on Soil Water Dynamics in a Rainfed-Canola Cropping System

The HYDRUS (2D/3D) model was applied to investigate the probable effects of different subsurface drainage systems on the soil water dynamics under a rainfed-canola cropping system in paddy fields. Field experiments were conducted during two rainfed-canola growing seasons on the subsurface-drained paddy fields of the Sari Agricultural Sciences and Natural Resources University, Mazandaran Province, northern Iran. A drainage pilot consisting of subsurface drainage systems with different drain depths and spacings was designed. Canola was cultivated as the second crop after the rice harvest. Measurements of the groundwater table depth and drain discharge were taken during the growing seasons. The performance of the HYDRUS-2D model during the calibration and validation phases was evaluated using the model efficiency (EF), root mean square error (RMSE), normalized root mean square error (NRMSE) and mean bias error (MBE) measures. Based on the criteria indices (MBE=0.01-0.17cm, RMSE=0.05-1.02 and EF=0.84-0.96 for drainage fluxes, and MBE=0.01-0.63, RMSE=0.34-5.54 and EF=0.89-0.99 for groundwater table depths), the model was capable of predicting drainage fluxes as well as groundwater table depths. The simulation results demonstrated that HYDRUS (2D/3D) is a powerful tool for proposing optimal scenario to achieve sustainable shallow aquifers in subsurface-drained paddy fields during winter cropping. Copyright (c) 2017 John Wiley & Sons, Ltd. Resume Le modele HYDRUS (2D/3D) a ete applique pour etudier les effets probables de differents systemes de drainage souterrain sur la dynamique de l'eau du sol dans le systeme de culture du colza pluvial implante dans des rizieres. Des experiences sur le terrain ont ete menees pendant deux saisons de culture de colza pluvial dans les rizieres drainees de l'Universite des sciences agricoles et des ressources naturelles de Sari, province de Mazandaran, dans le nord de l'Iran. Un pilote de drainage etait compose de drains enterres a differentes profondeurs et differents ecartements. Le colza a ete cultive comme deuxieme recolte apres la recolte de riz. Des mesures de la profondeur de nappe et de debit ont ete effectuees pendant les saisons de croissance. La performance du modele HYDRUS-2D lors des phases d'etalonnage et de validation a ete evaluee a l'aide de l'efficacite du modele (EF), de l'erreur carree moyenne (RMSE), de l'erreur carree moyenne normalisee (NRMSE) et des mesures de l'erreur de biais moyen (MBE). Sur la base des indices de criteres (MBE=0.01-0.17cm, RMSE=0.05-1.02 et EF=0.84-0.96 pour les flux de drainage, et MBE=0.01-0.63, RMSE=0.34-5.54 et EF=0.89-0.99 pour profondeurs de nappe), le modele etait capable de predire les flux de drainage ainsi que les profondeurs de la nappe phreatique. Les resultats de la simulation ont demontre que la gestion de la nappe souterraine peut etre une strategie efficace pour maintenir les aquiferes a faible profondeur dans les rizieres drainees pendant la culture hivernale. Copyright (c) 2017 John Wiley & Sons, Ltd.