Evaluation of ensemble stream flow prediction (ESP) (case study: Roud Zard drainage basin)

One of the most important information which is effective in desirable utilization of water resources is the information related to predicting the future streamflow. Forecast uncertainty is one of the main issues in the utilization of forecast information to manage water resources system. The ensemble stream flow prediction (ESP) is one of the methods in considering the forecasts uncertainty. The goal of this research is to develop and evaluated monthly ESP for Roud Zard (River) basin in Iran. Two models have been used to generate ensemble traces: 1-developing a conceptual river flow model (CRFM), 2-Combining the CRFM model with the adaptive fuzzy inference system (ANFIS) and development of a hybrid model. In continuation the Event bias correction method is employed on generated ESPs and the results of the three predictions have been compared. The results show that the precision of stream flow simulation model had considerable effects in the quality of the ensemble prediction as far as average ranked probability score in hybrid model that had higher precision than CRFM model has decreased from.52 to.38 and the months with suitable prediction (positive skills score) increased from 83 to 115 months and in case of employing bias correction, there will be an increase to 119 months from the total 182 months of similar period.