本文利用球面距离的Ripley'K函数,分析了全球2009年甲型H1N1流感大流行早期疫情的点空间分布模式。同时,通过对比2000—2008年甲型流感病例数据,分析不同纬度国家2009年甲型H1NI流感新增病例数的时间序列特征及其与国家入境人数的相关性。结果表明,2009年甲型H1N1流感大流行早期疫情呈聚类分布,其L函数值最大值区间与65个全球城市的最大值区间相同。78.500的病例分布在全球城市周围600km半径内。时间序列特征总体上类似于历年甲型流感,但是北回归线以北部分国家在6、7月非甲型流感流行季节仍有大量病例出现。并且北回归线以北国家冬季暴发集中在第45周到第48周之间,早于历年甲型流感流行时间。进一步分析认为,全球城市是本次流感国际传播网络的关键节点。国际旅行是流感传播的重要途径,并在本次流感大流行前期主导着流感跨国传播方向。同时不同纬度的环境条件对2009年甲型H1N1流感大流行有重要影响。
In March of 2009, a novel swine-origin influenza A(H1N1) virus was first discovered in Mexico and quickly spread to over 200 countries in less than two years. However, limited research has been con- ducted on the characterization of the global spatio-temporal transmission of the pandemic. Applying Ripley's K function based on the spherical distances, we analyzed spatial pattern of the outbreaks of the H1N1 pandemic from March 15, 2009 to June 9, 2009. Compared with other type A influenza occurred during 2000- 2008, the 2009 H1N1 influenza showed generally similar temporal trend, but marked difference when we broke down the outbreak data of each country along the latitude. To look into the differences, we further associated the number of weekly cases of the H1N1 influenza with national arrivals through customs. Results show that the 2009 H1N1 influenza in early period was spatially clustered. The maxi-mum value of the function L was identical to that of the 65 global cities, within which 79 percent of the outbreaks were distributed within a radius of 600 km. In addition, the correlation coefficients show that the highest positive correlation (r = 0. 7, p =. 002) between national arrivals and weekly influenza cases lied in the 19th week. These findings suggest that global cities are the key nodes of the network which dis-seminates international travels, hence the viruses in the early period of the pandemic. It was found that the seasonal environmental factors also have impact on the influenza pandemic through applying time se- ries analysis. Unexpectedly, some countries in the northern temperate zone reported more confirmed hu-man cases in June and July when was thought not to be suitable for the transmission of the ipfluenza. In the meantime, the winter peaks of cases for the countries that lie to the north of the tropic of cancer are clustered around the period between the 45th week and the 48th week, which is earlier than the common type A influenza season. It might partially due to the lack of immunit