Comparison of BP and LSTM Neural Network for Hydrologic Forecasting of a Small Watershed in Fujian
2020
CUI Wei | GU Ranhao | CHEN Benyue | WANG Wen
LSTM (long short-term memory) neural network is a type of recurrent neural network with feedback connections, which can learn the state characteristics between time series data. So it is very suitable for rainfall-runoff forecasting. According to the hourly rainfall and runoff data of the Duli Hydrological Station in Yanshouxi River Basin of Fujian, this paper establishes the BP neural network and LSTM neural network by the modular modeling method, avoids the local optimization in the model training by the method of ensemble prediction mean, and conducts the rolling forecasting of hourly runoff within 1 to 24 hour by the two neural network models. The results show that the overall forecasting performance of the LSTM model is better than that of BP model, and the rolling forecasting accuracy of the LSTM model drops much slower than that of BP model. The Nash efficiency coefficient of 1~24 hourly forecasting is 0.968~ 0.740, which can meet the requirements for short-term flood forecasting accuracy.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par Directory of Open Access Journals
Découvrez la collection de ce fournisseur de données dans AGRIS
