Quantitative assessment of ecological flow in the Yellow River under changing environments

Abstract

Studying the streamflow characteristics of the Yellow River mainstem under changing environments is of great significance to the management and sustainable development and utilization of water resources in its basin. In this paper, a long short-term memory (LSTM) model is used to restore the flow characteristics of the mainstream of the Yellow River under natural conditions, and the range of variation approach (RVA) and the nonparametric kernel density estimation (KDE) method are combined to quantitatively assess the impact of the changing environment on the streamflow. The study shows that: (1) The hydrological variability of the Yellow River occurred in 1985, and the degree of variability ranged from 26–58%, which is moderate. (2) The annual ecological flow value of the Yellow River ranges from 560 ~ 1001 m3/s, and the annual average ecological flow guarantee is 43%; (3) Through LSTM simulation (NSE > 0.7, R2 > 0.8), it is concluded that the ecological flow assurance in the natural situation of the Yellow River is higher than the measured value, which is mainly affected by human activities, and its contribution rate is more than 52%. This shows that the river ecosystem of the mainstem of the Yellow River is relatively unstable and needs further management.