Monte Carlo approach to identification of the composition of stratospheric aerosols from infrared solar occultation measurements

We describe an inversion method for determining the composition, density, and size of stratospheric clouds and aerosols by satellite remote sensing. The method, which combines linear least-squares minimization and Monte Carlo techniques, is tested with pure synthetic IR spectra. The synthetic spectral data are constructed to mimic mid-IR spectra recorded by the Improved Limb Atmospheric Spectrometer (ILAS-I and ILAS-II) instruments, which operate in the solar occultation mode and record numerous polar stratospheric cloud events. The advantages and limitations of the proposed technique are discussed. In brief we find that stratospheric aerosol in the size range from 0.5 to 4.0 (mu)m can be retrieved to an accuracy of 30%. We also show that the chemical composition of common stratospheric aerosols can be determined, whereas identification of their phases from mid-IR satellite remote-sensing data alone appears to be questionable.