Multiangle photopolarimetric aerosol retrievals in the vicinity of clouds: Synthetic study based on a large eddy simulation

We investigate the effect of cloud contamination and 3-D radiative transfer effects on aerosol retrievals from multiangle photopolarimetric measurements in the vicinity of clouds. To this end multiangle, multiwavelength photopolarimetric observations are simulated using a 3-D radiative transfer model for scenes with realistic cloud properties, based on a large eddy simulation. Spatial resolutions of 2 × 2, 4 × 4, and 6 × 6 km² have been considered. It is found that a goodness-of-fit criterion efficiently filters out cloud contamination. However, it does not filter out all scenes that are affected by 3-D radiative effects, resulting in small biases in the retrieved aerosol optical thickness (AOT) and single-scattering albedo (SSA). We also found that measurements at higher spatial resolution (2 × 2 km²) do not result in retrievals closer to clouds compared to measurements at coarser spatial resolutions (4 × 4 and 6 × 6 km²). If cloud parameters are fitted simultaneously with aerosol parameters using a 1-D radiative transfer model and the Independent Pixel Approximation, more successful retrievals are obtained in partially cloudy scenes and in the vicinity of clouds. This effect is most apparent at 6 × 6 km² and only marginal at 2 × 2 km² resolution. The retrieved aerosol AOT and SSA from the simultaneous aerosol and cloud retrievals still have a small bias, like the aerosol-only retrievals. We conclude that in order to substantially improve aerosol retrievals in the vicinity of clouds, a retrieval algorithm is needed that takes into account 3-D ratransfer effects.