论文部分内容阅读
Salinization is a gradual process that should be monitored. Modelling is a suitable altative technique that saves time and cost for the field monitoring. But the performance of the models should be evaluated using the measured data. Therefore, the aim of this study was to evaluate and compare the SALTMED and HYDRUS-1D models using the measured soil water content, soil salinity and wheat yield data under different levels of saline irrigation water and groundwater depth. The field experiment was conducted in 2013 and in this research three controlled groundwater depths, i.e., 60 (CD60), 80 (CD80) and 100 (CD100) cm and two salinity levels of irrigation water, i.e., 4 (EC4) and 8 (EC8) dS/m were used in a complete randomized design with three replications. Soil water content and soil salinity were measured in soil profile and compared with the predicted values by the SALTMED and HYDRUS-1D models. Calibrations of the SALTMED and HYDRUS-1D models were carried out using the measured data under EC4-CD100 treatment and the data of the other treatments were used for validation. The statistical parameters including normalized root mean square error (NRMSE) and degree of agreement (d) showed that the values for predicting soil water content and soil salinity were more accurate in the HYDRUS-1D model than in the SALTMED model. The NRMSE and d values of the HYDRUS-1D model were 9.6％ and 0.64 for the predicted soil water content and 6.2％ and 0.98 for the predicted soil salinity, respectively. These indices of the SALTMED model were 10.6％ and 0.81 for the predicted soil water content and 11.0％ and 0.97 for the predicted soil salinity, respectively. According to the NRMSE and d values for the predicted wheat yield (9.8％ and 0.91, respectively) and dry matter (2.9％ and 0.99, respectively), we concluded that the SALTMED model predicted the wheat yield and dry matter accurately.