以典型半干旱区大同市1996~2011年实测日相对湿度资料剔除确定性成分、跳跃成分、周期成分得到的平稳随机成分序列进行自相关和偏相关分析,对比AR(P)模型建立了季节性AR(1)预报模型.验证模型表明,预报模型精度符合要求,可对未来大同市相对湿度进行短期预报,也可用于半干旱区月平均相对湿度的统计特征分析和短期预报,进而促进地区农业发展.
Taking the daily measured relative humidity 52 years' data in Datong city (a typical semi-arid area) as material, based on the hydrological sequence of analysis,the characteristics of autocorrelation and partial correlation analysis were studied with the composition of stationary random sequence by eliminating uncertainty component, jump, cycle component, and seasonal forecast model A R (1) was established through the comparison of A R (P) models. The results showed that the forecast model accuracy met the requirements, which could be used for the short-term forecast of the relative humidity in Datong city in the future, and could also be used for on semi-arid areas. The average relative humidity statistical characteristic analysis and short-term forecast is benefit for promoting regional agricultural development.