Prediction of streamflow based on the long-term response of streamflow to climatic factors in the source region of the Yellow River

Abstract

Study region: The source region of the Yellow River (SRYE) is located in the eastern part of the Tibetan Plateau is a major water production and water conservation area for the Yellow River. Study focus: This study aims to investigate the correlations between streamflow and meteorological factors/ocean in the SRYE from 1960 to 2018 using wavelet analysis. The effects of meteorological factors/ocean signals on streamflow were calculated using the Partial Least Squares-Structural Equation Model (PLS-SEM). Furthermore, climate factors with strong correlation with streamflow were used as inputs to random forest (RF) and multiple linear regression (MLR) models to predict monthly streamflow. New hydrological insights: Meteorological factors showed stronger correlation with streamflow compared to ocean signals, explaining 79.3% of the streamflow variation and much higher than ocean signals (0.1%). Among the meteorological factors, precipitation had the largest direct effect on streamflow (P 0.05), which together explained 78% of the streamflow variability. Temperature and relative humidity are two important factors that indirectly influenced streamflow through potential evapotranspiration (P < 0.01). Finally, precipitation, relative humidity, and minimum temperature were chosen as streamflow predictors in the SRYE. The RF showed a better performance in predicting long-term monthly streamflow than the MLR under complex climate-hydrological system.

Keywords

• Streamflow prediction
• Partial least squares-structural equation
• Wavelet coherence
• Random Forest
• Source region of the Yellow River