Quantifying uncertainty sources in runoff change attribution based on the Budyko framework.

[Display omitted] • The impact of the uncertainties on the attribution of runoff changes was quantified. • A method was improved to analyze non-monotonic trends in wet, normal, and dry years. • The breakpoint was the main uncertainty source for runoff change attribution. Quantifying the contributions of climate change (CC) and human activities (HAs), which are important drivers of the hydrological cycle, to runoff change by CC and HAs under uncertainty is crucial for water resource management, especially in karst basins. In this study, we adopted the climate elasticity method to assess the impacts of CC and HAs on runoff change in Chengbi River Basin during 1970–2019. Furthermore, the main sources of uncertainty (Budyko formula, breakpoint, and underlying surface characteristics) in runoff change attribution were quantified via analysis of variance method. The improved innovative trend analysis based on the Fisher optimal partition method was first used for the non-monotonic trend analysis. The groundwater storage change obtained through the abcd model was used to obtain the effective precipitation for the Budyko hypothesis. The results indicated that the runoff and precipitation exhibited an increasing trend in wet and normal years, whereas they exhibited a decreasing trend in dry years after 1995 compared with 1970–1994. Furthermore, potential evapotranspiration exhibited no significant trend. The climate elasticity method based on the Budyko formula was effective, and the Choudhury-Yang formula was found to be more suitable than the Fu formula for the basin. The impact of water storage change was not negligible, and water storage change needs to be considered in runoff change attribution. Effective precipitation played a dominant role in runoff change, with a mean contribution rate of 87.85%, whereas the mean contribution rates of potential evapotranspiration and underlying surface characteristics were 22.55% and −10.40%, respectively. The breakpoint was the main source of uncertainties in runoff change attribution and demonstrated the highest contribution to the uncertainties (85.93%). The simultaneous breakpoint of the runoff coefficient was suggested as a reference to determine the confident breakpoint of runoff. The findings of this study are expected to provide a scientific basis for water resource management and contribute to the development of the climate elasticity method. [ABSTRACT FROM AUTHOR]