Calibration and validation the SWAT model in the simulation of runoff and sediment in Shahr Chai of Miyaneh

Proper management of catchments is one of the most important ways to make optimal use of water and soil resources. In our country, most of the catchments, especially the mountainous catchments, do not have enough hydrometric and sedimentation stations. This fact makes any development and management plans difficult. Hydrologists and water resources researchers have come up with various solutions but none of them have been completely successful (Roustamiyan et al., 1999; 588 & Shaygan et al., 2011; 2). On the other hand, the limited methods of measurement in hydrology and the need to have a method to generalize the available statistics to areas without statistics or places where measurement is not possible. Also, simulating future hydrological changes is one of the main reasons for hydrological simulation (Beven & Binley, 2001; 46). The ability of the SWAT model to simulate the complex hydrological processes of watersheds in the GIS environment distinguishes this model from integrated models in which larger user units are the basis of operation. 

The study area is part of the Qizil Üzan River. Shahar Chai basin along with other rivers such as Zanjan Chai, Aydughmush, and Qaranquchay is one of the sub-branches of Qizil Üzan that flows northwest and north of the basin and joins near the Myaneh city. The data used in this study include a digital elevation model of 1: 10000 of mapping organization, land use, soil, precipitation, minimum and maximum temperature, wind speed, solar radiation, relative daily humidity of synoptic stations, Sarab, Heris, Bostan Abad, Charoymaq and Runoff and the sediment of Shahar Chai station. In the watershed of Shahar Chai, a soil map prepared by natural resources of East Azerbaijan province has been used. Based on these maps, 21 soil texture classes can be distinguished in the whole basin. Also, based on the land use map, six land use classes in the area were identified.

After parameterization and data entry, the simulation was performed for 20 years from January 1, 2000, to December 31, 2019, based on a monthly time step. To determine the degree of sensitivity of flow parameters in the SWAT model, sensitivity analysis was performed using the SUFI-2 method for 25 selected runoff parameters and 15 selected sedimentation parameters. Using validation results to remove parameters that are less sensitive from the calibration process, it is decided that finally the parameters with lower sensitivity were removed and 13 parameters for runoff and 7 parameters for sediment were selected that were more sensitive. The calibration model for runoff and sediment was done in one step with 1500 simulations in three replications. The calibration process ends when, based on the objective function, the coefficients required for evaluation are acceptable. According to the obtained results, all the evaluation criteria of the model in the simulation of runoff and sediment are allowed.

Examination of the results of the SUFI-2 method in the Shahar Chai basin showed that, based on the evaluation criteria of the coefficients of determination and Nash-Sutcliffe, both in the calibration and validation stages, it has good results in this basin. But they can't predict peak discharge and sediments well. To better determine the performance of the model, first of all, it is recommended that the statistics of stations and numerous and daily runoff and sedimentation measurements should be used instead of monthly, if any, in a basin, to compare their results. Secondly, to achieve the desired results, this model should be used in comparison with other simulation models in this basin and adjacent basins.