山东省玉米气象产量分离方法的多重比较分析

摘要

玉米是山东省主要的粮食作物之一，探究气象条件对山东省玉米产量的影响具有重要意义。本文采用山东省玉米生产相关数据，利用5a 滑动平均法、双指数平滑法、回归分析法、HP 滤波法对山东省玉米趋势产量进行拟合，并分离出相应的气象产量。利用 SPSS 相关分析和方差分析，对四种方法拟合趋势产量的显著程度和分离气象产量的准确程度进行比较。结果表明，四种方法均能较好地拟合趋势产量，但其分离气象产量的准确程度存在差异，准确性大小顺序为 HP 滤波分析＞回归分析＞5a 滑动平均法＞双指数平滑法。%Corn is one of the major food crops in Shandong Province.It has great significance to explore the influences of meteorological conditions on the yield of corn in Shandong Province.In this paper,the relat-ed data of corn production in Shandong Province were used to fit the trend yield by 5a sliding average method, double exponential smoothing method,regression analysis and HP filtering method respectively,and then the meteorological yield was separated out.By using SPSS correlation and variance analysis,the significant degree of the four methods for fitting the trend yield and the accuracy of separating meteorological yield were com-pared.The results showed that all the four methods could better fit the trend yield,but the accuracy of meteor-ological yield separation was different.The accuracy of the four methods was ordered as HP filtering analysis >regression analysis >5a sliding average method >double exponential smoothing method.