Application of Stochastic Simulation Approach in Daily Reference Evapotranspiration Forecasting

Reference evapotranspiration (ET₀) affected by many uncertain complicated atmospheric variables is an important parameter of hydrological models of water balance study. Base on daily meteorological data from 1955 to 2006 in Gaoyou irrigation district, FAO56 Penman-Menteith method was used to calculate daily reference evapotranspiration. Through Kendall rank correlation test, the calculated ET₀ has significant increasing trend and seasonal fluctuation. Additionally the daily ET₀ series apart of tendency and seasonal factor, there is obvious cyclic fluctuation in the remained series. At the same time taken into account of the effectiveness of uncertain factors, daily ET₀ of long series can be divided into four factors (seasonal factor, tendency factor, cyclic factor and stochastic factor). Moreover monthly weather correction coefficient in Gaoyou irrigation district classified according to sunshine hours was considered as an input coefficient to predict 2007 and 2008 daily ET₀. Fourier series model was used to compare with the stochastic model. The results indicated that the stochastic model can well reflect the relationship between predicted daily ET₀ and calculated value and the model is effective to improve the accuracy of forecasting. The stochastic model is a proposed model to provide data for providing useful data for hydrological models and assessing hydrological impact of changing climate conditions.