摘要
本文根据南京市六个空气污染浓度监测站的2001年6月至2005年7月的SO2污染浓度监测资料分析南京市SO2污染浓度的时空分布变化特征,结果为6个站浓度值有明显的季节变化特征,且在冬季容易出现南北相反的分布特征。针对目前常用的回归预报方法在选取气象要素时没有考虑其互相之间的相关性的缺点,本文提出了一种建立在EOF展开基础上的首先使预报因子正交化,再与逐步回归方程结合并且资料逐日更新的变系数的新型统计预报模型,经过实际预报检验,预报准确率比较高,有很好的应用效果。
By studying the spatial and temporal change of the SO2 pollutant density data obtained from six observation stations in Nanjing, the result indicates that the density has seasonal change and regional distribution character. Because the currently used statistic forecasting methods of SO2 density can＇t take into account the correlation among forecasting factors, we provide a new method to build forecasting model with empirical orthogonal function（EOF） and the stepwise regression analysis method. During the forecast experiment, the result of new forecasting model is more accurate, and this method can be well used in SO2 forecast.
出处
《气象科学》
CSCD
北大核心
2006年第4期422-426,共5页
Journal of the Meteorological Sciences
基金
江苏省科委项目(编号:BS2005072)资助
