In this work we demonstrate a method to derive the concentration of aerosol components from the spectral measurements of AOD (aerosol optical depth) and single scattering albedo along with their size distribution and extinction profile available from AERONET (Aerosol Robotic Network) and MPLNET (Micro-pulse Lidar Network) stations. The technique involves finding the best combination of aerosol concentration by minimizing differences between measured and calculated values of aerosol parameters such as AOD, single scattering albedo, and size distribution. We applied this technique over selected sites in three different regions of the United States (West coast, Great Plains, and North-East). Our results are then compared with the measured concentration of aerosol components available from IMPROVE (Interagency Monitoring of Protected Visual Environments) and EPA (Environmental Protection Agency) network, as well as two different versions of the GFDL (Geophysical Fluid Dynamics Laboratory) General Circulation Model AM2 with online and offline aerosols. In general, concentrations retrieved by our technique compare well with the ground-based measurements, but there are some discrepancies possibly due to the inherent differences in temporal and spatial scales of data averaging or some of the assumptions made in our study. Over continental North America, the online version of AM2 appears to overestimate sulfate concentration approximately by a factor of two and underestimate organic carbon by nearly the same amount. Results of our sensitivity study show that the errors in the retrieval of black carbon and sulfate concentrations could be as high as 100% when there is a large bias of ∼0.05 in the reference values of single scattering albedo under high AOD (≥0.5 at 0.44 μm) conditions. Knowledge on the vertical distribution of aerosols is crucial for an accurate retrieval of surface concentration of aerosols. We also determine the composition and concentration of elevated aerosol layers using this technique.