RAIN-FALL MODELING - AN APPLICATION OF BAYESIAN FORECASTING

The rainfall-runoff modeling is very useful for forecasting purposes. A good methodology for forecasting the future stream flow is a key req uirement for designers and operators of water resources systems. A com promise between conceptual and classical time series modeling is appli ed to model the relationship between rainfall and runoff. The dynamic nonlinear model is composed of a probability distribution describing t he observation, a link function relating its mean to the so called sta te parameters and a system of equations defining the evolution of thes e parameters. Its Bayesian nature permits to take into account subject ive information: making forward intervention, defining monitoring sche mes and introducing smoothing facilities. An application using the dat a of Fartura river's basin is reported. The assessment of the prior di stribution is discussed and the predictive performance of the linear a nd the non-linear models is reported.