Analysis Of Future Runoff Changes And Uncertainty In The Yalong River Basin

At the same time that climate change has caused extreme climate events such as global temperature rises and droughts and floods,the impact on the hydrological cycle and water resources of the basin is also inestimable,and the security of water resources has restricted national economic and social development.Therefore,studying the hydrological response process of river basins under future climate change conditions is extremely important for the rational planning and utilization of local water resources and sustainable economic development.The research area of this paper is the Yalong River Basin,which is an important energy base for the development of western China and the transmission of electricity from the west to the east.Therefore,it is of great practical significance to conduct research on the runoff response of this region under future climate change.On the basis of reviewing domestic and foreign research progress,this paper couples downscaling models and distributed hydrological models to predict the runoff process of the study area under future climate changes.The following conclusions are obtained:(1)The SWAT hydrological model has good applicability in the Yalong River Basin.The most sensitive parameters for runoff simulation results in the study area are: soil wet bulk density(SOL＿BD),main channel Manning coefficient(CH＿N2),and soil saturation conductivity(SOL＿K).The model is simulated on a daily scale.The NSE and R~2 of most stations can reach 0.8 and above,and the percentage deviations are basically within 10%.(2)Using the large-scale forecast factors of the four global climate models(GCM)of Can ESM,CNRM,GFDL,and MIROC,an ASD downscaling model suitable for the study area is built and four models’ simulation of temperature and precipitation are output in three future emission scenarios(RCP2.6,RCP4.5 and RCP8.5).The results show that precipitation will increase in the future,and with the increase of greenhouse gas emission concentration,the increase in precipitation will be greater;the trend of temperature varies depending on the selected GCM.(3)The down-scaled output of future meteorological elements is coupled with the calibrated SWAT model to obtain the future runoff change process in the study area.The results show that the simulated runoff values of the four climate models in the three emission scenarios have all increased compared with the observed conditions in the base period;the 2090 S of each climate model under the RCP8.5 scenario is the period with the largest total runoff and the largest increase,while the smallest period is the 2030 S under RCP2.6;in addition,the runoff of each model in futural January to March will be reduced from the reference period,and the runoff in the flood season will be greatly increased.The largest increase period of runoff is June.(4)The Bayesian model averaging method(BMA)was used to collectively simulate the output runoff of the four models in the reference period,and the BMA integrated runoff forecast values and uncertainty intervals for the three future emission scenarios were also given.The results show that compared with the single GCM model,the BMA model does not only improve the accuracy of runoff forecasting,but also derives the runoff uncertainty interval with better properties,which improves the reliability of the simulation.