Discussion on the medium and long-term hydrological forecasting method in dry seasons of the Pearl River Basin

The medium and long-term runoff forecast is one of the important parts of the water regulation. At present, there are 6 main methods applied to the medium and long-term hydrological forecasting in the Pearl River Basin in dry season, which are the historical evolution method, mean generation function, periodic mean superposition analysis, Markov forecasting model, multivariate stepwise regression and the random forest model. This paper summarizes all the medium and long-term hydrological forecasting methods in the the Pearl River basin in the past, introduces their basic concepts and model principles, and further discusses and analyzes various methods in combination with the actual situation to improve the accuracy of medium and long-term hydrological forecasting.