مقایسه کارایی مدل AquaCrop و Saltmed در شبیه‌سازی عملکرد گندم و شوری خاک در حمیدیه خوزستان

Because evaluating and measuring efficiency in different irrigation managements is time-consuming with high cost, models can be applied for different irrigation scenarios and could overcome these problems. Variation Crop models are appropriate tools for investigating irrigation management and their effect on plant production. The purpose of this study was to evaluate the efficiency of two models, including AquaCrop and Saltmed, for simulation of yield and soil salinity variation. The accuracy of the results of the simulation models depends on the accuracy of the required data of the model. When the measurement and determination of the input data are valid, then the applicability of model will be evaluated in different conditions after calibration and validation. There are necessary parameters for applying Aquacrop and Saltmed models: details of crop growth stages, root effective depth, sowing and harvest date, harvest index, crop coefficients in different growth stages, duration of wheat phonological growth stages, initial soil salinity in different layers and some irrigation data including depth and time of water application. In this regard, the two models of AquaCrop and Saltmed have been investigated, which are among the functional models with high ability and efficiency to simulate crop yield and soil salinity performance.
This research was conducted in Ramseh region of Hamidieh, Khuzestan, Iran. The latitude and longitude are 33º4'N and 47º41'E, respectively. In this regard, three pilots of 10 hectares were selected and in each of these parts, three pilot trials with an area of 2000 square meters were considered for evaluation and measurement. In the first year, using traditional irrigation management of farmers and using the initial salinity levels, soil physical and chemical characteristics compared to the simulation of two models, Aquacrop and Saltmed. These models can simulate crop yield, Biomass and soil moisture variation in saline and non-saline condition. In this investigation, Aquacrop version No. 4(2012) and Saltmed No. 3-04-02(2015) have been used for crop production and soil salinity simulation. During irrigation season, reference evapotranspiration has been determined using climatic data of Ahvaz synoptic station. Gathered data included: minimum and maximum temperature, minimum and maximum relative humidity, precipitation, sunshine hours and wind speed. Also during the growing season, the volume of irrigation water, irrigation hours, water use efficiency and root depth were determined. At the end of the growing season, direct measurement of selected farms in three replications was carried out to determine the wheat grain yield and biomass along with soil moisture changes in the soil profile. According to the measured data, two models of Saltmed and Aquacrop were calibrated for the first year of cultivation. Results of calibration of the first year were evaluated for verifying the results of the two models in the second year. The standard error (SE), Root mean square error (RMSE), Normalized root mean square error (NRMSE), correlation coefficient (R²), mean bias error (MBE), and model efficiency (EF) were used to determine the accuracy of the models.
The results of the first year showed that the normalized error of root mean square index for grain yield and biomass values in the AquaCrop model were 4 and 5%, respectively, and these values for the Saltmed model, were 8 and 9%, respectively. This statistical index for simulation of soil salinity with Saltmed model was about 18% and in AquaCrop model was 53%. Therefore, the Saltmed model is more consistent with the measured data of soil salinity. Also, the results of the second year that were used for validation showed that the normalized error of root mean square index for yield and biomass in the AquaCrop model was 4% and 4% respectively, and these corresponding values for Saltmed model were 22 and 14% respectively. Also, simulated of soil salinity results in the second year with Saltmed model showed that this model with normalized error of mean square error(26%) was more accurate for soil salinity simulation than AquaCrop model with normalized error of mean square error equal to 47%; therefore, the AquaCrop model has a better performance than Saltmed model for yield simulation. Also, the performance of Saltmed model has high efficiency and accuracy in simulating soil salinity than AquaCrop model. So based on the goal of projects and precision of simulation data, these individual models can be applied.