摘要
This study focuses on the analysis of uncertainties in hydrologic series and their impact on water resources management decisions. The methodology is based on Monte Carlo simulation, where relevant parameters on the rainfall-runoff processes of a hydrologic model were sampled from suitable probability distributions in order to analyse the probability distribution of output variables in a water resources system model. The procedure was applied to three water resources systems in the Duero basin (Spain). Results showed that the performances of the water resources systems are very sensitive to uncertainties in the model input. The deviation from the reliabilities or cumulated supply deficits obtained in the water resources systems compared to the reference inflows was much larger when the perturbations were unfavourable than when they were favourable. This suggests that the uncertainty analysis may have important implications when using these performance measures of the water resources system in decision making, particularly in dry years. The uncertainty on hydrologic variables may change dramatically our assessment of the performance of a water resources system and thus hydrologic inputs should be evaluated with extreme care.