Estimating ground-level PM2.5 using aerosol optical depth determined from satellite remote sensing

We assess the relationship of ground-level fine particulate matter (PM2.5) concentrations for 2000-2001 measured as part of the Canadian National Air Pollution Surveillance (NAPS) network and the U. S. Air Quality System (AQS), versus remote-sensed PM2.5 determined from aerosol optical depths (AOD) measured by the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Multiangle Imaging Spectroradiometer (MISR) satellite instruments. A global chemical transport model (GEOS-CHEM) is used to simulate the factors affecting the relation between AOD and PM2.5. AERONET AOD is used to evaluate the method (r=0.71, N=48, slope=0.69). We find significant spatial variation of the annual mean ground-based measurements with PM2.5 determined from MODIS (r=0.69, N=199, slope=0.82) and MISR (r=0.58, N=199, slope=0.57). Excluding California significantly increases the respective slopes and correlations. The relative vertical profile of aerosol extinction is the most important factor affecting the spatial relationship between satellite and surface measurements of PM2.5; neglecting this parameter would reduce the spatial correlation to 0.36. In contrast, temporal variation in AOD is the most influential parameter affecting the temporal relationship between satellite and surface measurements of PM2.5; neglecting daily variation in this parameter would decrease the correlation in eastern North America from 0.5-0.8 to less than 0.2. Other simulated aerosol properties, such as effective radius and extinction efficiency have a minor role temporally, but do influence the spatial correlation. Global mapping of PM2.5 from both MODIS and MISR reveals annual mean concentrations of 40-50 ug/m(3) over northern India and China.