基于时间序列模型的降雨量预测分析

Rainfall prediction based on time series model

为了解济南市未来降雨的变化情况,以济南市1959～2015年降雨量数据为研究对象运用SPSS软件中时间序列模型ARIMA对其进行了建模;拟合1959～2015年的降雨量数据及预测未来5年的降雨量。ARIMA模型预测结果表明了济南市2016～2020年的年降雨量依次为676.5,635.5,689.8,630.7,663.3mm,5年的年平均降雨量为659.2mm,这与过去57年的多年平均降雨量较为接近,可以依此推测济南市未来5年出现干旱及洪涝灾害的可能性较小。可以看出,影响ARIMA模型预测结果的因素有很多,为了更加精准地预测降雨量,应当考虑多种要素并结合当地具体环境,建立符合当地的降水量预测模型。

In order to promote using of ARIMA model in production and life,the rainfall data from 1959 to 2015 in Jinan City were used as the research object to model the time series model ARIMA in SPSS software,and the rainfall in the next five years was fitted and predicted.The results of ARIMA model show that the annual rainfall in Jinan City is 676.5mm,635.5mm,689.8mm,630.7mm and 663.3mm respectively,and the annual average rainfall is 659.2mm in 5 years.This is the same as that of the past 57 years.The average annual rainfall is relatively close,you can infer that the next five years,Jinan City,drought and floods are less likely.There are many factors that affect the prediction results of the ARIMA model.In order to predict the rainfall more accurately,a variety of factors should be considered and combined with the local specific environment to establish a local precipitation forecasting model.