This study presents an ensemble stream flow simulation with probability of occurrence accounting for errors of the GR4J turned to the Senegal River upper the Manantali dam. Through this approach, probability is associated to simulation at each time, with step and reliability of issues depending on probability scores. Results including the reliability of the results permit decision-makers to judge the reliability of their simulations. Past errors pattern of the model is used to perturb the model for issuing ensemble scenarios rather than classical single deterministic. Basic and reverse Box-Cox transformation is carried out to allow treatment of the errors through weighted multivariate Gaussian distribution. Statistic errors are then used to perturb the hydrological model to obtain ensemble issues (RAW-Ens). Furthermore, a post-processing method (affine kernel dressing) is performed to produce a dressed ensemble (AKD-Ens). Ensembles are evaluated at deterministic and probabilistic scale. Diagrams (attribute and ROC) are also used for this purpose. Evaluating methods reveal through scores that the system is reliable and that dressing method (AKD) improves quality of the raw ensemble drawn from the perturbed model.
Key words: Senegal River, Bafing, Manantali dam, Ensemble simulation, Kernel dressing.
Copyright © 2019 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0