Multimodel assessments of human and climate impacts on mean annual streamflow in China.

Human activities, as well as climate change, have had increasing impacts on natural hydrological systems, particularly streamflow. However, quantitative assessments of these impacts are lacking on large scales. In this study, we use the simulations from six global hydrological models driven by three meteorological forcings to investigate direct human impact (DHI) and climate change impact on streamflow in China. Results show that, in the sub-periods of 1971-1990 and 1991-2010, one-fifth to one-third of mean annual streamflow (MAF) reduced due to DHI in northern basins and much smaller (<4%) MAF reduced in southern basins. From 1971-1990 to 1991-2010, total MAF changes range from -13% to 10% across basins, wherein the relative contributions of DHI change and climate change show distinct spatial patterns. DHI change caused decreases in MAF in 70% of river segments, but climate change dominated the total MAF changes in 88% of river segments of China. In most northern basins, climate change results in changes of -9% to 18% of MAF, while DHI change results in decreases of 2% to 8% in MAF. In contrast with the impacts of climate change that may increase or decrease streamflow, DHI change almost always contributes to decreases in MAF over time, wherein water withdrawals are supposed to be the major impact on streamflow. This quantitative assessment can be a reference for attribution of streamflow changes at large scales despite uncertainty remains. We highlight the significant DHI in northern basins and the necessity to modulate DHI through improved water management towards a better adaptation to future climate change. [ABSTRACT FROM AUTHOR]