Quantitative Assessment of Ecological Flow in the Yellow River Under Changing Environments

ABSTRACT Studying the streamflow characteristics of the Yellow River mainstem under changing environmental conditions is crucial for the management and sustainable development of water resources within its basin. This research employs a long short‐term memory (LSTM) model to restore the flow characteristics of the Yellow River's mainstream under natural conditions. Additionally, the range of variation approach (RVA) and nonparametric kernel density estimation (KDE) method are integrated to quantitatively assess the impact of environmental changes on streamflow. The findings indicate that: (1) Hydrological variability in the Yellow River was observed in 1985, with a degree of variability ranging from 26% to 58%, classified as moderate. (2) The annual ecological flow value of the Yellow River is 560–1001 m 3 /s, and the average annual ecological flow assurance is 43%. (3) Based on LSTM simulation results (NSE > 0.7, R 2 > 0.8), it is concluded that the ecological flow assurance under natural conditions in the Yellow River exceeds the measured values, primarily due to human activities, which contribute over 52% to this discrepancy. These results suggest that the river ecosystem of the Yellow River's mainstem is relatively unstable and requires further management.