Can We Rely on Satellite Visible/Infrared Microphysical Retrievals of Boundary Layer Clouds in Partially Cloudy Scenes? Implications for Climate Research

Abstract This study addresses the longstanding question of the reliability of gridded visible/infrared satellite cloud properties in partially cloudy scenes. By using in‐situ cloud probes and airborne Research Scanning Polarimeter (RSP) observations, we analyze bias changes in satellite retrievals from the Spinning Enhanced Visible Infra‐Red Imager (SEVIRI) geostationary sensor during the ORACLES campaign. Biases in cloud optical depth ( τ ) and droplet effective radius ( r e ) modestly change for cloud area fraction greater than 35%. The agreement between SEVIRI and RSP r e substantially improves when the retrievals are averaged after removing pixels with τ < 3.0, yielding biases indistinguishable from overcast scenes. In addition, satellite and RSP show an excellent agreement for closed‐ and open‐cell stratocumulus clouds, showing that the satellite retrievals capture spatial changes of r e , and confirming that satellites can faithfully reproduce real physical features for optically thick and partially cloudy scenes. We demonstrate that a simple methodology can minimize uncertainties in satellite‐based climate studies.