中文摘要：

利用空气质量模式Model-3/CMAQ及京津冀地区高分辨率排放源清单,针对有代表性的污染时段（2012年2月7～16日）,设置了5种不同时刻的减排方案（在污染峰值提前4d、提前3d、提前2d、提前1d及当天减排）,对比在同样的减排比例下,不同时刻开始减排的效果差异.研究发现,提前采取减排控制措施比污染峰值当天开始减排对降低PM2.5浓度的影响更为明显,而且提前采取应急减排的时间越早,PM2.5浓度下降越明显.提前1d、2d、3d减排海淀站和城六区峰值浓度下降率分别为23%和22%、31%和30%、39%和38%,均明显高于当天减排的峰值浓度下降率10%和9%.但随着提前天数的增加,PM2.5峰值浓度进一步下降的幅度越来越小,减排效益较之前显著降低.提前4d减排海淀站和城六区峰值浓度下降率分别为40%和39%,提前4d减排和提前3d减排对降低污染峰值日PM2.5浓度的效果已没有太大差别.同时针对另一个污染时段（2012年1月11～20日）进行了相似的敏感性试验,得出了类似的结论.因此,针对某些污染事件的应急减排,综合考虑减排成本和减排效果,根据气象条件的预报,在可能引起重污染事件的不利气象条件来临时提前2～3d采取减排措施效果最好,既能有效降低PM2.5浓度,也可以避免因盲目长时间减排造成的成本过大.

英文摘要：

The Models-3Community Multi-scale Air Quality（CMAQ） Modeling system with a high resolution inventory data over Beijing-Tianjin-Hebei area was used to investigate the effects on PM2.5 concentrations over Beijing of emission-sources reduction, with the same reduction rate at 5different time points： 4 days, 3 days, 2 days, 1 day and 0 day in advance of the most polluted day. Simulations were made for a representative air pollution episode（Feb 7^th～16^th, 2012）, in which Feb 13th was found to be the most polluted day. The results show that the PM2.5 concentration was likely to decline more significantly if emission-sources reduction measures were taken before the most polluted day than were taken on the most polluted day. In addition, the earlier emission-sources reduction measures were taken, the more significantly the PM2.5 concentration would decline. Reducing emission-sources 1 day, 2 days, 3 days ahead of the most polluted day led to declination of the peak value of PM2.5 concentration at the Haidian station by 23%, 31%, and 39%, and in urban Beijing by 22%, 30%, and 38%, respectively. However, as the number of days ahead of the most polluted day（Feb 13th） to take reduction measures increased further, the additional decrease of the peak PM2.5 concentration became smaller, thus the emission-sources reduction benefits became less effective. The peak PM2.5 concentration would decrease by 40% and 39% at Haidian station and urban Beijing if the reduction measures were taken 4 days before the most polluted day, which shows almost no improvement compared with those 3days in advance. Similar results were obtained in simulations for another pollution episode（Jan 11th～20th, 2012）. For controlling severe air pollution, both reduction costs and benefits should be considered. Our study indicates that the most effective way of emission-sources reduction is to take reduction actions 2～3days ahead of the possible severe pollution event, which can be obtained from meteorological condition prediction. In t