摘要
以PM2.5污染的空间分布为研究对象,改进传统Moran＇s I指数,以便适用于大样本数据的空间自相关分析。通过改进Moran＇s I指数推导出期望与方差对改进Moran＇s I指数进行统计意义上的显著性检验,从而对PM2.5污染情况进行空间自相关性的判断与评价,揭示PM2.5污染呈现的空间分布趋势。为说明改进Moran＇s I指数的实际应用价值,采用成都市区7个空气质量监测点（不包括灵岩寺监测点）PM2.5每小时数据,进行改进Moran＇s I指数的计算与分析,绘制改进全局指数时序图和改进局部Moran＇s I指数渲染图。结果表明,成都市区全局Moran＇s I指数的范围为0.728 7-0.998 5,即PM2.5污染整体呈现显著空间聚集趋势,从局部看,PM2.5污染的空间聚集和空间异质特性随着时间转移至不同监测点,地理位置邻近或环境类似的监测点有类似的分布趋势。
Regarding the spatial distribution of PMzs pollution as the research object, traditional Moran＇s I is improved to adapt spatial autocorrelation analysis in large sample data. The expectation and variance of improved Moran＇s I are derived and using them to test significance, thereby doing spatial autocorrelation analysis and evaluation of PM2.5 pollution and then distribution trends of PM2.5 pollution is revealed. To illustrate the practical value of improved Moran＇ s I, PM2.5 hourly data of 7 air quality monitoring sites in Chengdu urban district （except Lingyansi air quality monitoring site）are adopted to calculate and analysis of improved Moran＇s L Improved global Moran＇s I timing diagram and improved local Moran＇s I rendering diagram are drawn. The results show that global Moran＇s I range from 0.728 7 to 0.998 5 in Chengdu urban district, that is, PM2.5 pollution presents spatial aggregation significantly. From the local perspective, spatial aggregation and spatial heterogeneity characteristics of PM2.5 shift to different monitoring sites over time. Air quality monitoring sites of similar location and environment have similar distribution trends.
出处
《环境科学与技术》
CAS
CSCD
北大核心
2016年第9期187-193,共7页
Environmental Science & Technology
基金
四川省教育厅重点项目(2015ZA0030)
可视化计算与虚拟现实四川省重点实验室项目(KJ201410)
全国统计科学研究计划一般项目(2013LY028)
