Retrieval of the Optical Parameters of the Atmosphere from Satellite Radiative Observation

Data of remote multiangle measurements of reflected radiance are used for retrieval of the optical thickness, single scattering albedo and phase function parameter of cloudy and clear atmosphere. The method of perceptron neural network allows obtaining the surface albedo, optical thickness, single scattering albedo and phase function parameter from input values of multi-angle radiance and solar incident angle of clear pixels. All mentioned parameters were randomly varied on the base of statistical models of possible measured parameters variation. An analytical method of the retrieval is applied to remote observations of reflected radiance of cloudy pixels. The slight horizontal heterogeneity of cloud is approximately taken into account. The simultaneous retrieval of the optical thickness and single scattering albedo at every wavelength independently and without rough restrictions on parameters retrieved is the advantage comparing with earlier studies. The methodology of the asymmetry parameter retrieval is proposed.