2007~2016年北京天气分型与霾日的关联

采用COST733软件将北京地区2007~2016年的大气环流总体分为T1~T9种类型，研究其与霾日的关联性，并结合PM_2.5和臭氧地面观测，分析不同天气型对应的污染特征及气象参数分布规律.2007~2016年霾日发生概率21.5%，T4和T9型下霾日最多，T5和T8型最不利于霾日发生.9类天气型下霾日变化具有阶段性，2007~2012年（阶段一）霾日少且年际差异小，2013~2016年（阶段二）霾日增多.对9类天气型下霾日PM_2.5及臭氧变化进行分析，T1、T3、T4和T9型霾日多出现在秋冬季，PM_2.5日变化为逐时增加态势，4类天气型在第一阶段的白天有浓度波动增长形成的小峰值，但第二阶段减弱消失.大部分天气型的霾日，阶段二的PM_2.5浓度较阶段一降低，T7和T9型表现为增加，增幅分别为23.7%和3.9%.所有天气型霾日的臭氧日变化均为单峰型，峰值出现在下午，臭氧日均浓度最高为T8型.此外，阶段二与阶段一相比，T3、T5和T6型臭氧平均浓度增加，其中T5型增幅达到49.8%.将霾日与近地面气象要素关联分析，平均气温、风向、风速可以较好的解释臭氧浓度变化，而PM_2.5的变化特征不仅与气象要素相关，在一定程度上也体现了污染排放及区域联动减排的贡献，需两者结合才能更好探究PM_2.5浓度整体特征及细节变化.

Objective analysis on circulation types and its links to haze days over Beijing during 2007~2016

Nine weather types (T1 to T9) during 2007 to 2016 in Beijing were first identified with the objective weather classification approach in COST 733. The correlation between these weather types and haze days was then investigated, and the characteristics of air pollution and meteorological parameters under nine weather types were analyzed in combination with the variation of surface PM_2.5 and ozone. The overall occurrence probability of haze days was 21.5% in Beijing during 2007~2016, and haze days were the most frequented in T4 and T9. The variation of haze days in 9 weather types could be divided to two stages. During 2007~2012 (stage 1), haze days were fewer and the interannual change was not significant, while number of haze days increased during 2013~2016 (stage 2). After analyzing the variation of PM_2.5 and ozone concentrations under nine types, it was found that haze days in T1, T3, T4 and T9 mostly occurred in autumn and winter, and the PM_2.5 concentration increased hour by hour. On stage1, there were concentration fluctuations and a daily peak in the morning, which decreased and disappeared on stage2. The PM_2.5 concentrations on stage 2 in haze days were lower than that on stage 1 except for those under T7 and T9 types, which were 23.7% and 3.9% higher. The diurnal variation pattern of ozone had a single-peak in haze days in nine types, with maximum concentrations in afternoon. Daily average ozone concentration of T8 was the highest. In addition, mean ozone concentrations in T3, T5 and T6 were higher on stage 2 than those of stage 1, and the degree of increase in T5 was highest (49.8%). Correlation analysis between haze days and meteorological elements demonstrated that, temperature, wind direction and speed can explain the ozone variation well. While PM_2.5 variation was not only related with the local meteorological elements, but also reflected the emission characteristics and contributions of regional linkage emission reduction to some extent. Thus, a comprehensive consideration of all these factors was a better way to research the PM_2.5 characteristics.