Short-term Forecasts Of Streamflow In The Uk Based On A Novel Hybrid Artificial Intelligence Algorithm

This study proposes a novel ensemble model, combining a Deep Learning algorithm, Nonlinear AutoRegressive network with eXogenous inputs, and two Machine Learning algorithms, Multilayer Perceptron and Random Forest, for the short-term streamflow forecasting. The model was tested on 18 watercourses throughout the United Kingdom and outperformed simpler models, with values of R2 above 0.9 for several watercourses. The hybrid Machine Learning-Deep Learning model was also shown to be less affected by reductions in performance as the forecasting horizon increases, leading to reliable predictions even for 7-day forecasts.