Attribution of climate change and human activities to streamflow variations with a posterior distribution of hydrological simulations

[摘要] Hydrological simulations are a main method of quantifyingthe contribution rate (CR) of climate change (CC) and human activities (HAs)to watershed streamflow changes. However, the uncertainty of hydrologicalsimulations is rarely considered in current research. To fill this researchgap, based on the Soil and Water Assessment Tool (SWAT) model, in thisstudy, we propose a new framework to quantify the CR of CC and HAs based onthe posterior histogram distribution of hydrological simulations. In our newquantitative framework, the uncertainty of hydrological simulations is firstconsidered to quantify the impact of “equifinality for differentparameters”, which is common in hydrological simulations. The Lancang River(LR) basin in China, which has been greatly affected by HAs in the past 2 decades, is then selected as the study area. The global gridded monthlysectoral water use data set (GMSWU), coupled with the dead capacity data ofthe large reservoirs within the LR basin and the Budyko hypothesisframework, is used to compare the calculation result of the novel framework. The results show that (1) the annual streamflow at Yunjinghongstation in the Lancang River basin changed abruptly in 2005, which was mainly due to the construction of the Xiaowan hydropower station thatstarted in October 2004. The annual streamflow and annual mean temperaturetime series from 1961 to 2015 in the LR basin showed significant decreasing and increasing trends at the α =  0.01 significance level, respectively. The annual precipitation showed an insignificantdecreasing trend. (2) The results of quantitative analysis using the newframework showed that the reason for the decrease in the streamflow atYunjinghong station was 42.6 % due to CC, and the remaining 57.4 % wasdue to HAs, such as the construction of hydropower stations within the studyarea. (3) The comparison with the other two methods showed that the CR of CCcalculated by the Budyko framework and the GMSWU data was 37.2 % and 42.5 %, respectively, and the errors of the calculations of the newframework proposed in this study were within 5 %. Therefore, the newlyproposed framework, which considers the uncertainty of hydrologicalsimulations, can accurately quantify the CR of CC and HAs to streamflowchanges. (4) The quantitative results calculated by using the simulationresults with the largest Nash–Sutcliffe efficiency coefficient (NSE) indicated that CC was the dominant factor in streamflow reduction, which was in opposition to the calculation results of our new framework. In otherwords, our novel framework could effectively solve the calculation errorscaused by the “equifinality for different parameters” of hydrologicalsimulations. (5) The results of this case study also showed that thereduction in the streamflow in June and November was mainly caused bydecreased precipitation and increased evapotranspiration, while the changesin the streamflow in other months were mainly due to HAs such as theregulation of the constructed reservoirs. In general, the novel quantitativeframework that considers the uncertainty of hydrological simulationspresented in this study has validated an efficient alternative forquantifying the CR of CC and HAs to streamflow changes.

[学科分类] 妇产科学