Lee Harry F, Fei Jie, Chan Christopher Y S, Pei Qing, Jia Xin, Yue Ricci P H
Department of Geography and International Center for China Development Studies, The University of Hong Kong, Pokfulam Road, Hong Kong.
Centre for Historical Geographical Studies, Fudan University, Shanghai 200433, China.
Soc Sci Med. 2017 Feb;174:53-63. doi: 10.1016/j.socscimed.2016.12.020. Epub 2016 Dec 15.
This study seeks to provide further insight regarding the relationship of climate-epidemics in Chinese history through a multi-scalar analysis. Based on 5961 epidemic incidents in China during 1370-1909 CE we applied Ordinary Least Square regression and panel data regression to verify the climate-epidemic nexus over a range of spatial scales (country, macro region, and province). Results show that epidemic outbreaks were negatively correlated with the temperature in historical China at various geographic levels, while a stark reduction in the correlational strength was observed at lower geographic levels. Furthermore, cooling drove up epidemic outbreaks in northern and central China, where population pressure reached a clear threshold for amplifying the vulnerability of epidemic outbreaks to climate change. Our findings help to illustrate the modifiable areal unit and the uncertain geographic context problems in climate-epidemics research. Researchers need to consider the scale effect in the course of statistical analyses, which are currently predominantly conducted on a national/single scale; and also the importance of how the study area is delineated, an issue which is rarely discussed in the climate-epidemics literature. Future research may leverage our results and provide a cross-analysis with those derived from spatial analysis.
本研究旨在通过多尺度分析,进一步深入了解中国历史上气候与流行病之间的关系。基于公元1370年至1909年间中国的5961起疫情事件,我们应用普通最小二乘法回归和面板数据回归,在一系列空间尺度(国家、宏观区域和省份)上验证气候与流行病的联系。结果表明,在中国历史上,不同地理层面的疫情爆发与气温呈负相关,而在较低地理层面,相关强度明显降低。此外,气候变冷导致中国北方和中部的疫情爆发增加,在这些地区,人口压力达到了一个明显的阈值,放大了疫情爆发对气候变化的脆弱性。我们的研究结果有助于说明气候与流行病研究中可修改的地域单元和不确定的地理背景问题。研究人员在统计分析过程中需要考虑尺度效应,目前统计分析主要在国家/单一尺度上进行;还需要考虑研究区域划分方式的重要性,这一问题在气候与流行病文献中很少被讨论。未来的研究可以利用我们的结果,并与空间分析得出的结果进行交叉分析。