Evaluating the uncertainty of eight approaches for separating the impacts of climate change and human activities on streamflow

Abstract

Numerous studies have separated the impacts of climate change and human activities on streamflow. However, the uncertainties of various methods applied for separating such impacts are seldom studied. Here, we comprehensively investigated the uncertainty and causality of eight methods (two statistical methods, two sensitivity coefficient methods and four hydrological models) for separating the contributions of climate change and human activities to streamflow over 20 catchments from 1952 to 2016 within the Yellow River Basin. The ensemble mean of the eight methods indicates that human activities play a leading role in streamflow changes in most selected catchments. However, the contributions vary greatly across different methods. The contribution of human activities in change period II is greater than that in change period I in most catchments that have two change periods, showing a gradually increasing trend in the contribution of human activities. With the increase in the contribution of climate change (or human activities) to streamflow, the uncertainty first increases before the contribution reaches ~50% and then decreases. Our study suggests that it is useful to use multiple methods to reduce uncertainty in quantifying the impacts in catchments where climate change and human activity have similar influences.