A framework for attributing runoff changes based on a monthly water balance model: An assessment across China

During the last decades, significant changes in runoff (Q) have been reported in many regions, and attributing the changes is of great significance for water resource management. The conventionally used elasticity method based on the Budyko hypothesis neglects the impacts of climate seasonality on annual Q change (ΔQ), and it is not yet well understood how the climate change and anthropogenic activities influence seasonal Q. Therefore, we propose a framework based on the ABCD model to explicitly identify the effects of climate change and anthropogenic disturbance on annual and seasonal ΔQ, and further apply it in 191 catchments across China from 1960 to 2000. The trend in annual Q exhibits a significant (α= 0.05) decreasing trend in most northern catchments and increasing trend in some catchments of the lower reaches of Yangtze River and Southeast Basin. The trend in seasonal Q in the northern catchments shows decreasing trend during all seasons, while most in the southern catchments shows increasing trend especially in summer. Regarding the causes for annual ΔQ, climate change has positive and negative effects in 60 % and 40 % catchments, respectively, and is the dominant factor in the catchments of the Yangtze River Basin, Southeast Basin and Pearl River Basin. Human activities have positive and negative effects in 20 % and 80 % catchments, respectively, and are the dominant factor in north China. Precipitation is the dominant climatic driver in 72 % catchments. For the causes for seasonal ΔQ, climate change increases Q in southeastern catchments during all seasons, while it decreases Q in northern catchments during autumn and winter. Human activities decrease Q in more than half of the catchments except in winter. The climate seasonality cannot be ignored and our proposed framework is superior to the elasticity method in capturing the impacts of climate seasonality on ΔQ. The elasticity method causes more than 5 % deviation of the contribution ratio of climate change to ΔQ in 48 catchments. In addition, this framework is reliable on multi-annual timescales and provides important reference for water resource management.