Feasibility study of integral property retrieval for tropospheric aerosol from Raman lidar data using principal component analysis

Abstract

A method is introduced to derive integral properties of the aerosol size distribution, e.g., aerosol mass, from tropospheric multiwavelength Raman lidar aerosol extinction and backscatter data, using an adapted form of the principal component analysis (PCA) technique. Since the refractive index of general tropospheric aerosols is variable and aerosol types can vary within one profile, an inversion technique applied in the troposphere should account for varying aerosol refractive indices. Using PCA, if a sufficiently complete set of appropriate refractive index dependent kernels is used, no a priori information about the aerosol type is necessary for the inversion of integral properties. In principle, the refractive index itself can be retrieved, but this quantity is more sensitive to measurement errors than the various integral properties of the aerosol size distribution. Here, the PCA technique adapted for use in the troposphere is introduced, the refractive index information content of the kernel sets is investigated, and error analyses are presented. The technique is then applied to actual tropospheric Raman lidar measurements.