Sensitivity Analysis Of Water Resources In Arid Inland River Basin Based On SWAT Model

With the increase in population and the development of social economy, thedemand for water resources have increased significantly, which makes the prominentcontradiction between water supply and demand. In addition, global warming willexacerbate the shortage of water resources, especially in arid area. Therefore, thestudy of the impact of climate change on water resources, have great significance forwater resources regulation, rational development of water resources control program,and security of water resources.In recent years, the distributed hydrological model, SWAT, has been widelyused in many fields. This model is a powerful tool for the runoff simulation, responsesimulations to basin climate change, and also the contaminant transport simulation.Supported by the RS and GIS, the land use map and soil map were reclassified tomeet the requirements of the model input data. On this basis, we establish theappropriate soil database and meteorological databases.Based on the DEM, land use map, soil type map, the Kaidu River is divided into33sub-basins and197hydrological response units. We selected the period of1986-1990and1986-1990for calibration and validation, respectively.Both the monthly and daily runoff for Dashankou station was simulated. For themonthly calibration of1986-1990, the efficiency coefficient (Ens), the determinationcoefficient (R2), and the relative error(RE) is0.92,0.90,0.24%, and those coefficientsfor daily calibration is0.70,0.73,-1.27%, respectively. While for the validationperiod, the monthly coefficients for Ens, R2, and RE are0.89,0.92and0.19, andthe daily coefficients are0.7,0.73,11.58%, respectively. Overall SWAT model cansimulated the streamflow well in the Kaidu River. So the results are satisfied forsimulating the hydrological processes in the Kaidu River. The difference between thesimulated and observed data is very mall, but still within the permissible error range. This study set up five different climate change scenarios for increasedtemperature by0℃、0.5℃、1.0℃、1.5℃and2.0℃, respectively.The futuretemperature, precipitation, relative humidity and wind speed was simulated based ondownscaling model (STAR). Combined with the runoff of SWAT under scenariossimulation, we analyzed runoff changes in the Kaidu River. Overall, the runoff willdecrease in the future scenario, and the monthly runoff is less than the multi-yearobserved average. Among the five scenario, the monthly average runoff in scenario1is small, scenario V is larger, and other scenarios are medium. In the spring, only5%of the future runoff is greater than the runoff in the base time. In the summer,15.7%of the predicted runoff is greater than the runoff of the base time, whichindicted the future runoff will decrease significantly. The study can provide scientificbasis and technical support for water management and drought/flood prevention.