Periodic Overlapping Prediction Model and Its Three Improved Models of Hydrological Time Series

The periodicity is usually implied in the hydrological phenomenon in the nature. Based on the characteristics of hydrological time series, periodic overlapping prediction model is proposed. In this paper, periodic overlapping prediction model is introduced at first; then the insufficiency of periodic overlapping prediction model under certain conditions are pointed out based on the analysis of examples, and three improved models are put forward. The first improved model is trend - periodic overlapping combinatorial prediction model (TPOLC); the second improved model is amplitude-compacted periodic overlapping prediction model (AC-POL); the third improved model is dominant period inversely-deduced periodic overlapping prediction model (DPI-POL). These models are all improved based on the practical work, which can be extensively used to build the modeling of corresponding type of hydrological time series.