Prediction and Analysis of Saltwater Intrusion in Pearl River Estuary Based on Machine Learning

Abstract

The acceleration of urbanization has induced a sharp increase in urban water consumption, so the saltwater intrusion has stronger influences on domestic, industrial, and agricultural water use in estuarine areas. To enhance the water supply security of coastal cities, it is necessary to analyze and predict saltwater intrusion. This study aims to further investigate the influence of various factors, including estuary tide level, wind direction and speed, and upstream water flow, on the saltwater intrusion (chlorinity in the estuary) in the Pearl River estuary, providing scientific support for improving water supply security in coastal cities. The study determines the lag time of different influencing factors on saltwater intrusion through correlation coefficients and constructs prediction models for the chloride content at the estuaries of the Guangchang and Pinggang water pumping stations based on various machine learning methods to analyze the saltwater intrusion in the Modaomen waterway of the Pearl River estuary. The established models show a good performance in predicting salinity. According to the analysis of the importance of influencing factors, it is found that upstream water flow has the greatest influence on saltwater intrusion, followed by estuary tide level and wind direction and speed.

Keywords

• saltwater intrusion
• Pearl River Estuary
• machine learning