Retrieval of Physical Properties of Atmospheric Particles by Inversion via Regularization in the Limit of a Small Optical Data Set

Atmospheric aerosols, although only a minor constituent of the earth’s atmosphere, play an important role in many atmospheric processes. Due to their appreciable influence on the earth’s radiation budget, air quality, clouds and precipitation as well as the chemistry of the troposphere and stratosphere it is necessary to gather detailed information on their optical and physical properties. A multiple-wavelength lidar as well as a Raman lidar at the Institute provide optical particle information in terms of six backscatter coefficients and two extinction coefficients in the wavelength range from 0.355 to 1.064 μm on a vertical scale. A data-evaluation algorithm that uses the method of inversion via regularization has been specifically designed to retrieve physical properties from the given optical information. The physical parameters can be described by, e.g., the particle size distributions, the mean sizes derived from it, like the effective radius, the volume, surface-area, and number concentrations as well as the complex refractive index. Due to the low amount of available a priori information on the particle properties in combination with the small number of optical information that additionally include large measurement errors the main focus had been on the retrieval of the mean values. To control the quality of the regularization under these difficult conditions the method of generalized cross-validation is used as it does not require the knowledge of the underlying measurement errors nor the knowledge of the specific shape of the particle size distributions.