Modeling River Discharge using Deep Learning in the Ouémé catchment at Savè outlet (Benin, West Africa)

This paper presents a modeling approach based on Artificial Neural Networks (ANNs) in the Ouémé river catchment at Savè. To do this, we used precipitation data as input over period 1965 -2010 to simulate discharge study area by using two ANNs models such Long Short Term Memory (LSTM) and Recurrent Gate (GRU) models. Indeed, description of stochastic nature is better presented today than statistical We compared performance these different evaluation criteria. The predictions made show strong similarity between observed simulated flows. deep learning gave good results. calibration validation, Nash Sutcliffe Efficiency (NSE) coefficient determination (R²) are very close one (calibration: R²= 0.995, NSE= 0.991, RMSE= 0.18; validation: R² = 0.975, 0.971, 0.41). LSTM GRU confirms importance Intelligence hydrological phenomena for decision-making.