First Retrieval of Aerosol Vertical Profile With Passive Remote Sensing: Part 1. Development of Algorithm Theoretical Basis

Abstract This paper presents the first part of a two‐part study to develop a new algorithm to retrieve the aerosol vertical extinction profile using the hyperspectral measurements at ultraviolet bands, O 2 A‐band and B‐band, from the Tropospheric Monitoring Instrument (TROPOMI). We represent the aerosol vertical profile by the weighted sum of 3–5 most important EOFs (empirical orthogonal function, i.e., eigenvectors) from the principal component analysis (PCA) of the 15‐year record of aerosol extinction profiles from spaceborne lidar Cloud‐Aerosol Lidar with Orthogonal Polarization (CALIOP). Hence, the retrieval is simplified to derive 3–5 coefficients or weights of corresponding EOFs to capture the variation of aerosol vertical profiles. A new PCA module was developed in the Unified Linearized Vector Radiative Transfer Model (UNL‐VRTM) for calculating the Jacobians of top‐of‐atmosphere (TOA) reflectance with respect to the weights of EOFs, which is used to facilitate the optimal inversion of the EOF weights. The analytical Jacobian calculations are validated against the Jacobians computed from a finite difference method. The averaging kernel analysis for directly retrieving the aerosol extinction profiles from measurements of TROPOMI and high‐resolution metagrating spectropolarimeter for aerosol profiling was provided. Finally, the retrieval experiments with synthetic TROPOMI measurements generated by UNL‐VRTM were conducted to verify the self‐consistency and feasibility of the inversion algorithm on a theoretical basis.