中文摘要：

采集兰州市2005年1月1日至2007年12月31日的大气颗粒物（TSP、PM10、PM2.5、PM1.0），分析不同粒径颗粒物的污染特征及其与气象因素的相关关系，并采用时间序列半参数广义相加模型（GAM）分析了可吸入颗粒物与呼吸系统疾病和心脑血管疾病日入院人次的暴露-反应关系.结果显示，TSP和PM10的质量浓度在春季较高，PM2.5和PM1.0在冬季较高；气象因子与大气颗粒物有显著的相关关系；对于呼吸系统疾病来说，PM10、PM2.5和PM1.0每升高10 μg·m-3或10粒子数 m-3，入院危险分别增加0.052%、0.604%和0.652%；对于心脑血管疾病来说，PM10、PM2.5和PM1.0每升高10 μg·m-3或10粒子数 m-3入院危险分别增加0.046%、0.697%和0.935%.由此可见，兰州市不同粒径大气颗粒物均有不同程度的污染，PM10、PM2.5和PM1.0对呼吸系统疾病和心脑血管疾病入院人次均有一定影响，且PM10对呼吸系统疾病的影响较心脑血管疾病明显，而PM2.5和PM1.0则正好相反.

英文摘要：

Individual atmospheric particles including TSP, PM10, PM2.5 and PM1.0 were collected in Lanzhou from January 1, 2005 to December 31, 2007. The relationships between different sizes of atmospheric particulate matter and meteorological factors were analyzed. A semi-parametric Generalized Additive Model （GAM） was applied to analyze the exposure-response relationships between mass concentrations of inhalable particle and daily hospitalizations of respiratory, cardiovascular and cerebrovascular diseases. The results indicated that mass concentrations of TSP and PM10 were higher in summer, but PM2.5 and PM1.0 concentrations were higher in winter. Moreover, the meteorological factors were significantly correlated with atmospheric particles. In addition, respiratory hospitalization risk increased by 0.052%,0.604% and 0.652% with increase of 10 μg m-3 or 10 particle number m-3 in PM10,PM2.5 and PM1.0; hospitalization risk for cardiovascular and cerebrovascular diseases increased by 0.046%,0.697% and 0.935% with increase of 10 μg m-3or 10 particle number m-3 in PM10,PM2.5 and PM1.0 respectively. In conclusion, significant variation existed in pollution levels of atmospheric particles with different particle sizes in Lanzhou. Mass concentrations of PM10, PM2.5 and PM1.0 had some influences on hospitalizations of respiratory, cardiovascular and cerebrovascular diseases. Additionally the impact of PM10 on the respiratory diseases was more obvious than that on the cardiovascular and cerebrovascular diseases, while the impacts of PM2.5 and PM1.0 were contrary to that of PM10.