Quantitative Evaluation of Uncertainty in Predicting Future Runoff and Water Storage in a Watershed
The accuracy of the hydrological simulation is closely related to the multi-source uncertainty involved in the hydrological process,and its cumulative effect inevitably leads to further expansion of hydrological prediction uncertainty.Therefore,this paper takes the source region of the Yellow River as the research object and uses the maximum likelihood uncertainty estimation method(GLUE)to effectively identify the"same effect"parameter group of the Xin'anjiang mod-el.By coupling four future climate models,three climate change scenario data,and eight"different parameters and same effect"parameter sets under CMIP5,the impact of source uncertainty on watershed runoff and water storage was ex-plored.Finally,the analysis of variance(ANOVA)method was used to quantitatively separate the relative contributions of each source uncertainty to monthly scale watershed runoff and water storage.The results show that the precipitation,minimum and maximum temperature data obtained using the SDSM downscaling model can be better applied to the Yellow River basin,with correlation coefficients(R2)greater than 0.70 and root mean square error(RRMSE)less than 30% .The uncertainty of parameters and GCMs dominates the impact on watershed runoff,with the former contributing up to 0.98 relative to watershed water storage,and the interaction between multi-source uncertainties contributes more to pre-flood and post-flood than to non-flood period.The research results are of great significance for flood control and disas-ter reduction in river basins,as well as for reducing cognitive uncertainty in hydrological simulation processes.
multi-source uncertaintyrunoffwater storagehydrologic modelUpper Reaches of Yellow River
万方数据
维普
