Impact Of Climate Change On Runoff And Components Based On Semi-distributed Hydrological Model

Climate change has attracted global attention.Scientists,engineers and decision makers are interested in the hydrological responses to climate change.An exacerbated uneven distribution of precipitation in time and space due to climate change has led to frequent floods and droughts,which threaten life security of human beings and usually cause large economic losses.Therefore,quantifying the potential impacts of future climate changes on runoff is important to facilitate public decision making for more appropriate strategies and improved water resource management.Most studies of runoff are conventionally concentrated on total runoff.The accuracy of outputs is fairly well based on traditional performance function like Nash-Sutcliffe efficiency coefficient.However,various uncertainties,especially equifinality,make the investigation doubtful.This thesis took the Yarlung Zangbo River as a case study.Multiple methods in climate change,including global climate models,downscaling method,hydrological model,etc.Firstly,the Nyang River,one of the major tributaries of Yarlung Zangbo River,was selected to evaluate the feasibility of remote sensing data as well as to construct SWAT model.Based on the generated monthly runoff simulations,a preliminary runoff components research framework was established.Meanwhile,the developed simple batch-processing module was used to investigate historical runoff and corresponding components.Then,a SWAT framework was built over the entire Yarlung Zangbo River Basin.Monthly runoff was simulated,and accordingly a comprehensive runoff components research framework under changing climate was established.With the purpose of exploring runoff and corresponding components(groundwater,rainfall and meltwater induced runoff)under changing climate,a main stream batch-processing module were developed.Finally,SWAT model was further constructed to simulate daily runoff.A distributed batch-processing module at watershed scale was developed to study runoff and its components,namely groundwater,lateral runoff and surface runoff,under changing climate.The major findings of this study are summarized as follows:(1)In order to overcome data shortage in the Nyang River Basin,the feasibility of TRMM remote sensing data was evaluated.Then,historical runoff was reconstructed based on TRMM remote sensing data via SWAT model for the Nyang River Basin.A simple batch-processing module was developed based on source code of SWAT model to investigate groundwater,rainfall induced runoff and meltwater induced runoff.Hydrological signatures were utilized to verify the simulations.The results showed that TRMM remote sensing data was applicable to drive hydrological modelling in the study area.Both total runoff and corresponding components agreed well with gauge data and hydrological signatures.Furthermore,the preliminary runoff components research framework was suitable for the Nyang River Basin.(2)LARS-WG stochastic weather generator was employed to downscale outputs from five global climate models(GCM),namely BCC-CSM-1-1,CanESM2,CSIRO-Mk3-6-0,GISS-E2-R and HadGEM2-AO under RCP2.6 and RCP8.5 scenarios.The downscaled GCM data were then post processed by Schaake shuffle procedure to obtain daily precipitation and temperature timeseries in the far future.The developed main stream batch-processing program based on source code of SWAT was used to investigate simulated historical and future groundwater,rainfall-induced and snowmelt-induced runoff by SWAT at Nugesha,Yangcun and Nuxia stations.The results indicated that groundwater at all three stations would decrease while rainfall-induced runoff would increase.Snowmelt-induced runoff at Nugesha and Yangcun stations would increase whilst decrease at Nuxia station.(3)A daily SWAT model was established based on obtained historical and downscaled GCM data to simulate historical and future daily runoff.Components,including groundwater,lateral runoff and surface runoff,were studied by a developed watershed batch-processing program.Water engineering time series processing tool was utilized to screen the best parameters set and to verify simulations.Correlation length theory was adopted for hydrological simulations.The combined daily hydrological model and watershed batch-processing program showed that groundwater and lateral runoff at all three stations would decrease while surface runoff would increase.In addition,the time correlation lengths of the three components are likely to decrease in the future,and the update speed of the corresponding components will be accelerated.