Radiation measurements from satellites, aircraft and the ground have been successfully employed for characterizing radiative properties of aerosols, clouds, atmospheric gases, land, and ocean. One of the challenges is the development of a reliable procedure for inversion of the observations. This session is dedicated to unite works emphasizing the various aspects of numerical inversion that improve quantitatively and/or qualitatively the retrieved atmospheric products. The contributions are expected to address such important attributes of inversion as optimum accounting for errors in the data, inverting multi-source data, utilizing a priori information, inverse modeling and data assimilation, retrieval error estimations, and retrieval acceleration. Development of forward radiative transfer models to improve retrieval is also welcome. We encourage explorations of new retrieval concept and improved products such as aerosol types and profiles, surface particulate matter, trace gas, clouds, and land and ocean properties for existing and next generation of satellite missions and ground-based networks.