This paper presents a methodology for deriving aerosol climatology with artificial intelligence using lidar observations combined with atmospheric modeling and unsupervised deep learning. The method can be applied to lidar measurements and modeled data, or their combination, if observed and modeled data are consistent. The advantages of the method are: ability to automatically retrieve the optical properties of aerosols, repeatability and scalability.
Talianu et al. (Thu,) studied this question.