Multi-station Water Quality Prediction Considering Temporal and Spatial Correlation: Based On T-GCN

The water quality monitoring data reflects the ecological environment of the river and its changing trend. The accurate analysis and prediction of water quality is of great significance to improve the efficiency of river management and maintenance, ensure the safety and stability of the ecological environment of the river basin. The traditional water quality prediction model combines hydrodynamic theory and water quality change process mechanism, which can represent the overall change trend of water quality. However, this method is mostly aimed at a fixed watershed, with poor portability ability and difficult to accurately predict key areas. Although some new mathematical statistical methods and machine learning methods can achieve high accuracy in the prediction of water quality data from a single station, the spatial correlation characteristics between multiple stations are ignored in the comprehensive water quality prediction of the basin, resulting in a decline in accuracy. In this paper, based on the spatial distribution and graph structure characteristics of multiple water quality monitoring stations in the basin, the temporal graph convolutional neural network is used to analyse the time series characteristics of water quality, and realize the prediction of water quality data from multiple monitoring stations. The analysis of water quality monitoring data of a river located in the Guangdong Province of China proves that, compared with the commonly used model LSTM, the T-GCN model can better fit the spatial correlation characteristics and time series characteristics of water quality monitoring data, and achieve a more accurate water quality predictive analysis. Based on the water quality monitoring data of multiple stations, T-GCN model can reflect the changing trend of water quality in the entire basin, and accurately predict the water quality changing in specific locations. The research results can provide a more accurate and effective reference for river operation and management.