Study On Methods Of Long-range Hydrologic Forecast By Comparison In Shanghai

The long-range hydrological forecast is a boundary science of hydrology, meteorology, climatology and other relative sciences. Because of complexity of affecting factors and restriction of current conditions, the results of the long-range hydrological forecast are still not perfect. How to improve the forecast accuracy is the key problem concerned about by hydrologists.In this paper, the long-range hydrological forecast is studied and Shanghai water regime is selected as a case study. Emphasis is put on the comparison of different methods in order to improve forecast accuracy. The prediction of the annual highest tide stage at Huang-pu Park is very important, but in the past it is of little effect. Why? Perhaps it is upland water, astronomical bearing and typhoon that conduces to difficulties of forecasting the annual highest tide stage. In the paper, because sea temperature, Southern Oscillation index and sunspot number are also investigated, forecast accuracy is improved, especially in prediction of the annual highest tide stage.The paper lays emphasis on researchs of stepwise multiple regression, empirical orthogonal function, projection pursuit regression, in comparison with autoregression model, grey model (1,1), and threshold autoregressive model .Moreover, in the end weighted Markov chain model, Markov chain model and comparison of them are studied that refer to qualitative forecast...