Distributed modeling of soil water-salt dynamics and crop yields based on HYDRUS-EPIC model in Hetao Irrigation District

Abstract: Soil moisture and salinity are two key factors for crop production in arid irrigation districts. It is critical to modify soil water-salt dynamics and crop growth on a regional scale for the sustainable agriculture. In this paper, a distributed agro-hydrological model that well considers the spatial variability of soil and hydrological factors was developed to simulate soil water movement, solute transport and crop growth process on the regional scale. Jiefangzha Irrigation System (JIS) of the Hetao Irrigation District was selected as the study area. The JIS was divided into 201 homogeneous simulation units based on the combinations of weather-soil-crop-irrigation. In this way, the one-dimensional agro-hydrological model-HYDRUS-EPIC (HYDRUS-1D coupled with EPIC crop growth module), was used and expanded to the regional scale. Field experiments were conducted in 2012 and 2013. The dataset of soil moisture, soil solute concentration, leaf area index (LAI) and crop yield were collected at 40 monitoring points, and used for model calibration and validation. Simulated soil moisture and salinity concentration in the root zone showed good agreement with the measured values. During the calibration process, root mean square error (RMSE), mean relative error (MRE) and coefficient of determination (R2) for soil moisture were 0.03 cm3/cm3, 0.3% and 0.67, respectively. For salinity concentration, RMSE, MRE and R2 were 2.72 g/L, -13.5% and 0.53. LAI and crop yields were fitted well with the observations. MRE values for the estimated and measured LAI and crop yields were 1.0% and 1.1%, and R2 were both larger than 0.90 for these two items. During the validation process, RMSE, MRE, and R2 were 0.04 cm3/cm3, 2.6%, 0.57 for soil moisture, and 2.62 g/L, -4.5%, 0.51 for salinity concentration, respectively. And MRE and R2 were 9.1%, 0.88 for LAI, and -1.9%, 0.92 for crop yields. These results showed that the distributed agro-hydrological model was able to simulate the soil water flow, salt transport, and crop growth process in JIS with accuracy. The calibrated and validated model was then applied to predict spatial distribution of soil moisture, salinity concentration, crop evaporation and crop yields of the study area in present irrigation water management practices. Effective saturation and salinity concentration in the root zone were chosen to represent soil water and salinity stress on crop growth. Results showed that effective saturation ranged from 0.44 to 0.90 with an average of 0.7 for the JIS. In most areas, soil water could meet crop water consumption needs. In the areas where groundwater depth (GWD) was less than 1.3 m, root water uptake was limited due to waterlogging. The average salinity concentration in the root zone varied from 3.1 g/L in the northwest to 13.5 g/L in the northeast with an average of 6.4 g/L for the whole district. High soil salinity concentration limited crop production seriously. Corresponding to the spatial distribution of salinity concentration in the root zone, crop relative yield (ratio of actual yield and average yield of JIS) ranged from 0.33 to 1.33. The results suggested that for the northeastern part, where GWDs were larger than 2.0 m, more irrigation was needed for leaching salt. It was also better to plant more salt tolerant crops in these areas. In northwestern and southwestern parts, shallow groundwater levels intensified waterlogging or salinity accumulation problems. The study indicated that it is better to keep the groundwater depth not shallower than 1.3 m for maintaining the crop yields.