CNN-LSTM convolutional recurrent neural network hydrological forecast correction method based on grid rainfall information

The invention belongs to the technical field of watershed hydrological forecast error correction, and discloses a CNN-LSTM convolutional recurrent neural network hydrological forecast correction method based on grid rainfall information, and the hydrological forecast correction method is totally divided into three main steps. The method comprises the steps of hydrological element collection, CNN-LSTM convolutional recurrent neural network correction model construction and model verification and error correction. Collected data are used as input characteristics of the model, historical observation data and forecast data are used for training and determining a model structure, correlation among forecast influence factors, spatial-temporal characteristics and forecast errors is mined, the model is verified by using actual forecast rainfall runoff data, and intelligent correction of hydrological forecast errors is realized. The model has high precision, particularly in the flood season, the corrected forecast flo