中文摘要：

目的 探讨北京市2008—2009年大气PM10浓度的时空分布特征和人群暴露水平,分析北京市大气PM10浓度对居民呼吸系统疾病死亡的暴露-反应关系。方法 采用克里格插值模型对研究期间北京市大气PM10的日均浓度进行估计,采用时间序列的广义相加混合效应模型分析大气PM10浓度对居民呼吸系统疾病死亡的暴露-反应关系。结果 北京市2008—2009年大气PM10的日均浓度为118.6μg/m3,高于GB 3095—1996《环境空气质量标准》二级标准。研究期间大气PM10浓度呈现自北向南逐渐升高的空间分布规律。北京市大气PM10浓度每升高10μg/m3对呼吸系统疾病死亡的超额危险度为0.56%（95%CI：0.28%~0.83%）。引入多污染物模型后大气PM10浓度对呼吸系统疾病死亡的超额危险度略有减小,但仍有统计学意义（P〈0.05）。结论 采用克里格插值模型能够较为精确地估计北京市大气PM10浓度的空间分布状况;大气PM10浓度对北京市居民呼吸系统死亡存在一定的暴露-反应关系。

英文摘要：

Objective To explore the spatial-temporal characteristics of ambient particulate matter （PM10） mass concentration and formulate the exposure-response relationship between the ambient PM10 level and respiratory disease mortality of the exposed population in Beijing. Methods Kriging interpolation model was used to estimate the spatial-temporal characteristics and concentration of ambient PM10 geographically around Beijing and optimized based on the results of cross-validation. Generalized additive mixed model （GAMM） was used to examine the city-wide association between PM10 and respiratory disease mortality adjusting for the random effects of districts. Results The daily average concentration of ambient PM10 of Beijing was 118.6 μg/m3 during 2008 to 2009, higher than secondary standard of Ambient Air Quality Standard （AAQS, GB 3095-1996）. The spatial distribution of ambient PM10 concentration presented to increase gradually from the northern to the southern in Beijing. A 10 μg/m3 increase in the PM10 was associated with an increase of 0.56% （95%CI： 0.28%-0.83%） in respiratory mortality of the population in Beijing. Although, the odds ratio decreased after introducing the multiple pollutants models, it was also statistically significant （P〈0.05）. Conclusion Kriging interpolation models performs a good estimation of spatial distribution of the PM10 concentration across Beijing. The methods used in this study which combined the Kriging interpolation and GAMM may assess the exposure-response relationship between PM10 level and respiratory mortality in Beijing.