Effects of cloud variability on TROPOMI molecular and cloud property products

In order to guarantee high signal-to-noise ratios, the design of spectrometers aboard spaceborne platforms is a tradeoff between spectral and spatial resolution. Since molecular absorption is highly spectrally dependent, atmospheric composition instruments favor spectral over spatial resolution. The TROPOMI instrument aboard Sentinel 5 Precursor (S5P) will have a spatial resolution of about 7x7 km2 at nadir, which clearly outperforms the resolution of previous atmospheric missions (320x40 km2 for OME/ERS-2, ~120x30 km2 for SCIAMACHY/ENVISAT and 80x40 km2 for GOME-2/Metop-A,B). However, inside a TROPOMI ground pixel there may still be a considerable amount of unresolved cloud variability. In this work, we present a sensitivity study of measured reflectivities as a function of the unresolved cloud variability. For this purpose, we simulate virtual measurements of a TROPOMI-like instrument in cloudy scenes at different spatial resolutions by means of the three-dimensional (3D) radiative transfer model MoCaRT (Monte Carlo Radiative Transfer). The reference inside-pixel cloud microphysical properties are provided by the PArallelized Large-Eddy-Simulation Model (PALM) at a spatial resolution of 10x10 m2 in a 6.4x6.4 km2 domain. The retrieval algorithms of both, atmospheric molecules and cloud properties rely on one-dimensional radiative transfer (RT) models which do not account for neither cloud variability nor 3D RT effects. Moreover, the quality of atmospheric gas retrievals strongly depends on the accuracy of the cloud information. Accordingly, we first analyze the 3D effects on the cloud products by ingesting the 3D simulated spectra (around the oxygen A-band) to the ROCINN_CAL cloud algorithm. In a second step, the impact of the cloud variability on the ozone product is assessed (which also include the indirect impact on cloud properties).