Forecasting Solar and Wind Potentials in Southern Morocco Using Artificial Intelligence Models

Advanced climate vision to reduce the effects of climate change has become a necessity. This study aims to propose efficient methods for wind speed and solar irradiation forecasting based on previously available datasets using artificial intelligence (AI) models. Artificial neural network (ANN), recurrent neural network (RNN), and nonlinear auto-regressive exogenous (NARX) AI models were applied to predict the annual wind speed and global horizontal solar irradiation in southern Morocco. Three prestigious southern Moroccan sites with important solar and wind potentials were selected for this study: Dakhla, Laayoune, and Boujdour. The performance of these models was tested and validated using the available measured data. The best performance results were obtained using the ANN model with one hidden layer of 20 neurons, which presented minimum MSE errors at all evaluated sites compared to the RNN and NARX models.