Soil water dynamics on irrigated garlic and pepper crops using hydrus–1D model in the lake Tana-Basin, northwestern Ethiopia

Abstract

Soil water is an important variable in regulating and predicting hydrological process for optimal irrigation. Hydrus-1D was used to simulate soil water dynamics under overhead irrigation in Dengeshita watershed at the plot level. Experiments were carried out from October-February 2017/2018 and from March – June 2018. The treatments were conservation agriculture (CA) and conventional tillage (CT). Irrigation depth, crop phenology, meteorological and soil parameters were determined. Soil parameters were estimated using a K-nearest neighbor approach (KNN) pedotransfer functions for tropical soils and fitted using retention curve optimization program. Sensitivity analysis result showed saturated soil water content (θs), saturated hydraulic conductivity (Ks), and pore size distribution (n) were the most important parameters for the model. The model performance using measured soil water content (SWC) was good with R2 of (0.64–0.77) and errors; RMSE of 0.021–0.063 and ME of 0.0013–0.040. Based on overall evaluation, CA plots had higher average SWC (0.39–0.40 cm3.cm−3) than CT plots (0.36–0.37 cm3.cm−3). The average seasonal actual transpiration was lower for CT (88.76%) than CA (93.46%) plots due to higher evaporation loss (CT = 7.69% and CA = 1.15%); the difference is statistically insignificant. Seasonal deep percolation from CT and CA plots was 0.38% and 3.15% respectively. Therefore, CA was better than CT due to store more water for plants.