ABSTRACT
ABSTRACT: A group of probabilistic methods for hydrologic dam safety analysis is presented in this paper. The methods were proposed and validated in the MODEX project. The approach proposed in MODEX consisted on the generation of arbitrarily long series of basin response hydrographs in order to characterize the dam hydrologic forcing. Hydrographs can be generated from streamflow gauge observation through multivariate techniques or through use of multivariate stochastic rainfall generation models in conjunction with physically based distributed hydrological models. A pilot basin (Manzanares basin in Spain), where data availability allowed for proper probabilistic calibration of models,was selected. A Monte Carlo probabilistic simulation environment was developed integrating the hydrometeorological modeling chain. A computing platform was used to provide discharge time series that were used in the characterization of hydrographs for high return periods and applied to the hydrological dam safety assessment. Using the probabilistic framework, significant advances by accounting for uncertainties in the precipitation field and in processes governing the rainfall-runoff transformation were achieved. The scientific results of the project were applied to identify methodologies for analyzing hydrological dam safety under a probabilistic approach that can be widely implemented in the professional field. The determination of the real risk associated with existing infrastructure adds an objective criterion to the assessment and prioritization of investment on embankment dam protection.
