Application of seasonal time series model in the precipitation forecast

Abstract Precipitation is an important guiding standard for agricultural production; however, it is highly difficult to forecast due to random sequential and seasonal features. In this research on the historical data of time series, it is found that rainfall has a strong autocorrelation of seasonal characteristics in time series. Utilizing seasonal periodicity with a Seasonal Autoregressive and Moving Average (SARIMA) methodology we analyze the statistical data of precipitation, which is based on Shouguang city in Shandong statistical yearbooks over the 1996:1–2009:12 periods. In this paper the method for time series analysis and forecasting are proposed to study the model. The experimental results could achieve good prediction fitting degree. In this sense, the model is available for actual forecast warning in precipitation. Through the comparison of the model we had found the advantages of forecasting that can make full use of natural rainfall for corresponding areas and save underground water resources.