中文摘要：

针对传统地面监测手段稀疏获取PM2.5浓度的缺陷,该研究在利用中分辨率成像光谱仪(MODIS)遥感影像反演500m空间分辨率气溶胶光学厚度(AOD)的基础上,辅以人口密度、植被叶面积指数等地理特征要素,构建了研究区时均尺度PM2.5浓度土地利用回归(LUR)模型,并与经典克里格插值方法(OK)对比进行精度评价。结果表明:(1)研究区内部AOD存在明显空间分异,中心城区AOD值高于周边地区;(2)融合AOD要素的最优LUR模型相关系数可达0.51;(3)相比OK方法,LUR模型生成的PM2.5浓度图可客观反映研究区空气污染的空间变化。研究表明,融合遥感气溶胶数据的LUR模型可在短时间尺度准确模拟地面高空间分辨率PM2.5浓度。

英文摘要：

Relative to the defect of traditional monitoring method of PM2 .5 concentrations with sparse field campaigns, remote sensing has advantages of wide coverage and continuous observation in space.In this study,aerosol optical depth with 500m resolution was generated with the Moderate Resolution Imaging Spectroradionmeter (MODIS)image data,and the AOD was then used to build a land use regression (LUR)model for simulating the PM2 .5 concentrations in an hourly span,and the accuracy of the optimal modeling method was evaluated by comparing with the result of Ordinary Kriging (OK)interpolation. Then the optimal LUR modeling method was used to simulating and mapping PM2 .5 concentration in the study area.The results show that:(1)the AOD has obvious spatial differentiation of the study area and the values are higher in the downtown area than the surrounding area;(2)among of LUR models,the correlation coefficient of optimum LUR model can reach 0.51 by incorporating AOD data;(3)the optimum LUR model can well reveal the spatial variation of PM2 .5 with smaller relative errors than OK interpolation method.It is suggested that the LUR model by incorporating AOD data can accurately simulate PM2 .5 concentrations at high spatial resolution in the short time scale,and the proposed LUR model is an effective method of mapping PM2 .5 concentration with high spatio-temporal resolution under sparse ground conditions.