ABSTRACT
This Chapter analyses the uncertainty propagation due to the model. Comparisons between rainfall resolutions and hydrologic model resolutions serve as a guide for modelers and radar developers to know how much detail is necessary to archive a reliable solution in small watersheds in terms of flow prediction using ensembles. The chapter presents Predictability Limits Due to Up-scaling. Monte-Carlo method approximation is based on a large number of possible initial hydrologic states drawn up randomly from the Probabilistic distribution functions (PDF) of initial-condition uncertainty in the phase space. The stochastic dynamic simulation is constructed by a substantial amount of hydrologic simulations, repeatedly running the model is where the knowledge of the real PDF's are required. The goal of the evaluation is to determine significant differences between the mean of the ranked probability score for the highest resolution and the means for the other resolutions.
