Water flow and nitrate transfer simulations in rice cultivation under different irrigation and nitrogen fertilizer application managements by HYDRUS-2D model

Conducting field experiments to determine an optimal strategy for managing water consumption in paddy fields is quite time-consuming and expensive. This study applied HYDRUS-2D model to simulate fluctuations in water level and nitrate transfer over the 2 years of 2013 and 2014 at Rice Research Institute of Iran (Rasht). To monitor water-level fluctuations, a piezometer was installed in the plot. Three irrigation management regimes were considered: first, permanent water flooding or irrigating by as much as 1.3 times the amount of evaporation from the evaporation pan (I-1); second, applying irrigation treatment by one time the amount of evaporation from the evaporation pan (I-2); third, irrigating by 0.7 times the amount of evaporation from the evaporation pan (I-3) from the transplanting date to the end of irrigation period (10 days before harvest). Two fertilizing treatments were considered, i.e., C1 = 60 kg N ha(-1) and C2 = 60 kg N ha(-1) + 5000 kg compost ha(-1) equal to 160 kg N ha(-1). After model run, observed and simulated values were compared using EF (efficiency factor), RMSE (root-mean-squared error), and nRMSE (normalized root-mean-squared error) indices. The results of EF, RMSE, and nRMSE with respect to water-level changes at the calibration stage were 0.94, 5.39 cm, and 7.6%, respectively. At the validation stage, the results of EF, RMSE, and nRMSE were 0.96, 2.25 cm, and 3.2%, respectively. The nRMSE index showed the model's excellent performance in simulating water-level changes. In the case of nitrate transfer, at the calibration stage, the EF values at the two fertilizer levels of C1 and C2 were 0.60 and 0.85, and nRMSE values were 27 and 9.97%, respectively. At the validation stage, the EF values at the two levels of C1 and C2 were 0.60 and 0.99, and nRMSE values were 13.7 and 1.3%, respectively. Based on the values of nRMSE index, the performance of the model in simulating nitrate transfer was shown to vary in the medium-to-excellent class. Overall, the simulation results showed that the HYDRUS-2D model exhibited an appropriate ability to simulate water movement and nitrate transport either from chemical or compost sources; therefore, it can be used for irrigation and fertilizer management in the case of paddy fields to analyze different scenarios.