Estimating spatial variability of ground-level PM2.5 based on a satellite-derived aerosol optical depth product: Fuzhou, China

Estimating exposure to PM2.5 within urban areas has important implications for human health. Satellite remote sensing provides an effective means to retrieve spatial coverage of PM2.5 concentrations. Using the monitoring network established by the local government in 2014, this study developed a linear mixed effects model that integrates aerosol optical depth (AOD) measurements (spatial resolution: 3 km) from MODIS and meteorological data from GEOS-FP meteorological fields as predictors to derive daily estimations of ground-level PM2.5 concentrations in Fuzhou (SE China). A 10-fold cross validation method was employed to examine the performance of the mixed effects model. The cross validation yielded a R-2 value of 0.72, and a root mean square error value of 9.2 mu g/m(3) for the mixed effects model. Furthermore, the monthly/seasonal average PM2.5 estimated by the mixed effects model was highly correlated with those of in situ measurements. The results also revealed the spatial differences in the PM2.5 distribution across the study area, i.e., higher concentrations in the urban center and lower values in suburban to rural areas. The results suggest that the mixed effects model using MODIS 3 km AOD together with meteorological data could be effective for the estimation of PM2.5 concentrations in the Fuzhou area.