Isolating Cabannes and rotational Raman scattering to determine forward modeling error due to Rayleigh scattering in the Earth’s atmosphere

Molecular scattering in the Earth’s atmosphere is commonly represented in radiative transfer (RT) models using the Rayleigh formulation. We show this to be a monochromatic approximation that incurs non-negligible bias due to spectral structure. In reality, atmospheric molecular scattering consists of two distinct physical components: elastic Cabannes scattering and inelastic Raman scattering. vSmartMOM provides a physically consistent framework to resolve Cabannes and rotational Raman (RRS) scattering individually, enabling for the first time a study of their remote-sensing sensitivities. Our previous work focused on Raman scattering with little discussion of the Cabannes component. Here we present the formulation of Cabannes scattering by N 2 , O 2 and their combined effect in air, each described by its coupling with Rayleigh scattering and RRS. Rayleigh and RRS parameters derived from Placzek’s quantum-mechanical treatment of molecular polarizability are used to determine the exact Cabannes single scattering albedo and depolarization ratio critical to its accurate vector RT simulation. This is of particular importance for hyperspectral observations from the ultraviolet (UV) to the near-infrared (NIR). We focus on the O 2 A-band, where spectral structure causes significant Rayleigh bias despite weaker RRS relative to the UV. This has implications for GHG retrievals that widely use the O 2 A-band to constrain airmass, and for FLEX SIF retrievals aiming to determine the SIF spectrum from the A-band. We examine the dependence of the Rayleigh bias on surface brightness, pressure, viewing geometry, and post-cloud-screened residual aerosol, with a LuT-based recommendation to correct it.