中文摘要：

区域PM_（2.5）浓度影响因子及显著程度对区域PM_（2.5）浓度模拟和污染控制具有重要意义。该研究应用广义加性模型（GAM）建立模型分析2013年京津冀区域PM_（2.5）浓度与AOD、气象因子（相对湿度、温度、降雨量、大气压、风速）和土地利用类型（水体、林地、耕地、建设用地、裸地）之间的相关关系。结果表明,温度、大气压、AOD、林地、建设用地和裸地显著的影响PM_（2.5）浓度;且温度、AOD、裸地、林地与PM_（2.5）存在复杂相关关系,大气压、建设用地与PM_（2.5）浓度存在线性相关关系。GAM模型R~2为0.952,拟合结果与实测结果的线性回归方程系数为0.959,模型交叉验证后得到R2为0.792。结果表明,利用GAM能有效的识别区域PM_（2.5）浓度的影响因子,根据影响因子进行PM_（2.5）浓度拟合并得到可靠的拟合结果。

英文摘要：

The study of effect factors of regional PM_（2.5）concentration has important implications for regional PM_（2.5）simulation and pollution control. The correlation analysis between regional PM_（2.5）concentration of the year 2013 and its effect factors such as AOD,meteorological parameters such as relative humidity,temperature,precipitation,air pressure,wind speed and land use types such as water,forest,farm land,constructed land,bare land was conducted based on generalized additive model（GAM）. The results indicated that temperature,air pressure,AOD,forest,constructed land and bare land have a significant impact on PM_（2.5）concentration. There exists complex correlation between PM_（2.5）and temperature,AOD,bare land,forest and linear correlation between PM_（2.5）and air pressure,constructed land. R2 of GAM model is 0.952,the coefficients of linear regression between fitted and observed PM_（2.5）concentration is 0.959,and the cross validation R2 is 0.792,which indicated that GAM is effective enough to identify the effect factors of regional PM_（2.5）concentration and we are able to obtain reliable fitting results of PM_（2.5）concentration in regional PM_（2.5）concentration simulation with these parameters.