深圳市空气污染分区预报思路的建立

基于本市2005—2010年各区API日资料和污染日逐时资料、逐日气象要素资料、物理量因子，在对全市污染时空分布特征和各影响因素细致分析的基础上，运用典型天气形势和典型要素判定方法，在此基础上，进行气象要素的偏最小二乘回归提炼主要因子，建立污染天气预报思路，运用2011年重污染目的资料进行检验，结果表明：该种三级判定的方法预测污染小概率事件，吻合较好，为空气污染客观预报和高污染预警系统的建立提供依据。

Research on air pollution partition prediction method

Using API data, daily and hourly data of pollution days, daily meteorological data and physical factors of Shenzhen from 2005 to 2010, the prediction method of air pollution partition was analysed based on the spatial and temporal pollution distributions of Shenzhen and detailed influencing factors by typical elements and weather situation analyse method. Meanwhile, this paper uses partial least squares method and refine important meteorological elements to establish the pollution forecasting ideas. Through the test on data of 2011 heavy pollution days, the results show that this three level determination method has good effect on predicting the oecuranee of small probability pollution events and provides a good foundation for air pollution forecast and the establishment of early warning systems.