ABSTRACT
A hydrologic model was evaluated for its potential to perform real-time flood forecasting within the Mayaguez Bay drainage basin located in western Puerto Rico. Minimal run times, enhanced prediction skill, parameterization of variables and the understanding the dynamics of the system are issues that need to be faced to enhance flood prediction. In distributed models, the parameter values are physically based and the watershed is represented by grids, which approximates the parameter distribution and the initial conditions of the system. The modeler assigns the grid size resolution to the model, rainfall input scales and parameter values in a subjective way; subjective because the modeler has to select among various methods available for assigning grid point values, and each method can influence the hydrologic result of the model. Each parameter and input is spatially and temporally scale-dependent, probability distributions are not known a priori, and the implications, in terms of uncertainty propagation though the system, are well understood.
