On the retrieval of urban aerosol mass concentration by a 532 and 1064 nm LIDAR

Based on the hypothesis of a monomodal, lognormal size distribution, the un certainty affecting the humid-mass retrieval from LIDAR data was estimated by considering our ignorance of the distribution width to be a source of er ror. The mass to backscatter ratio and its uncertainty were computed for si x accumulation-mode aerosol models as a function of the backscatter angstro m coefficient (alpha) and of the relative humidity (RH). A mass to backscat ter uncertainly of less than +/- 30% was obtained for all six models. We co mputed the mass and simulated the expected LIDAR backscatter at 532 and 106 4 nm for a test data set of 14 "real-world" multimodal size distributions o btained from the literature. The possible presence of 0-20%-50% water-insol uble compounds in each aerosol mode was assumed. An urban-type accumulation mode and 10 different coarse mode compositions were considered, including dust-like aerosols. The aerosol mass concentration was derived by fitting t he simulated LIDAR data at 532 and 1064 nm with a monomodal distribution of urban aerosols of "unknown" width. The relative over- or underestimation o f the mass with respect to the real aerosol mass was expressed in terms of alpha and RH For the 10 coarse aerosol types. The LIDAR-derived mass turned out to be underestimated by 0 - 15% in the case of (NH4)(2)SO4, NaCl, mari time, and H2SO4 coarse aerosols. In the case coarse dust aerosols, the rang e of underestimation was wider (0-30%). Absorbing aerosols showed a maximum underestimation of 40-50%. (C) 2000 Elsevier Science Ltd. All rights reser ved.