Using MODIS derived aerosol optical depth to estimate ground-level PM(2.5) concentrations over Turkey

WOS: 000481603900022Satellite based particulate matter (PM) pollution monitoring on a regional basis is of importance due in part to the adverse health effects of PM. In this study, Moderate Resolution Imaging Spectroradiometer (MODIS) derived aerosol optical depth (AOD) data at 3 km and 10 km resolutions from both Terra and Aqua satellites were used, in conjunction with the surface in situ data, to improve the regional distribution of ground-level PM2.5 over Turkey. Five years (2011-2015) of heating season's (15th October to 14th May) in situ PM2.5 measurements from 7 monitoring stations in Ankara and 3 years (2013-2015) of the same data from 13 monitoring stations in Marmara Region were used. Linear and non-linear regression models were used to find the relationship between PM2.5 and AOD data. To improve the correlations between PM2.5 and AOD, the data points affected by free tropospheric long-range transport were removed from the analysis via back trajectory modeling analysis since long-range transport affects AOD more readily than surface PM2.5 data. Using non-linear models with the addition of meteorological parameters such as height of planetary boundary layer, surface temperature and surface wind speed improved the correlations significantly. The best non-linear model can explain 61% (n = 37, R-2 = 0.61, p < 0.001, RMSE = 0.337 mu g/m(3)) of PM2.5 variations at the Edirne Kesan site. It was found that Terra worked better than Aqua. Furthermore, 10-km aerosol products gave better correlations with PM2.5 as compared to the 3-km products. With the aid of spatiotemporal model, PM(2.5 )distribution maps are created for the first time for Turkey.Scientific and Technological Research Council of Turkey (TUBITAK)Turkiye Bilimsel ve Teknolojik Arastirma Kurumu (TUBITAK) [BIDEB-2219]; NASA ACMAP programThe authors would like to acknowledge the MODIS aerosol team for providing the satellite aerosol products. O. Z. greatly acknowledges the financial support from the Scientific and Technological Research Council of Turkey (TUBITAK) BIDEB-2219 Postdoctoral Research program. Y. W. was supported by the NASA ACMAP program. We are thankful to Dr. Guray Dogan for his support in back trajectory modeling