Application Of Factorial Design And Bayesian Probability Forecast Method On Hydrological Model On Xiangxi River Watershed

As the global climate changes and extreme weather frequently happens in recent years, the hydrological forecast plays a more prominent role in the national life. In the field of hydrological model forecast, there are many uncertainties within various models. As to uncertainty within models, the results tend to have big errors. With the advantage of ability to analyse relationship between parameters and the dependent variable as well as relationships between parameters, factorial design method is often used in engineering field. In this paper, HYMOD method is applied into xiang river basin, and the runoff is simulated. Then parameters of HYMOD is analysed through the factorial design method. At the same time, this paper established the bayesian probability hydrological forecasting system, which is used in Xiangxi River. Projection pursuit regression (PPR) model is used for long-term basin precipitation forecast, while hydrologic uncertainty processor (HUP) is used to deal with uncertainty of hydrological model, Finally the uncertainty of hydrological forecast is described in the form of probability distribution. After applying the bayesian probability hydrological forecasting system in xiang river valley, the results show that the bayesian probability hydrological forecast model obtained better results, compared with deterministic runoff forecast and other forecast models, for it is not only beneficial for decision makers to consider uncertainty of the model, and it can improve the precision of runoff forecast.