Climatology of marine low-cloud morphology types, their climatological transitions, cloud radiative and physical properties, and environment

A deep learning model is trained to classify marine low clouds into six mesoscale cloud morphology (MCM) types: solid stratus, closed-cellular convection, disorganized cellular convection, open-cellular convection, clustered cumulus, and suppressed cumulus. We present global maps of their frequency of occurrence and variances. Spatial distributions of each MCM type have clear regional features that are associated with large-scale circulation patterns and thermodynamic conditions. Low cloud MCM types undergo systematic climatological evolutions and transitions over five stratocumulus regions as the air mass moves from coastal region to the tropics. The characteristics of the transitions and evolutions share common features but also have regional signatures. Compared to subtropical stratocumulus regions, characters of cloud type evolution are slightly different over regions that are dominated by cold-air outbreaks although both share similar evolution paths. We also found that mesoscale mean cloud properties such as cloudy scene albedo, cloud optical depth, cloud fraction, droplet size, cloud top height and liquid water path show systematic dependence on the MCM type and strong intra-type variations. Systematic changes in cloud properties among MCM types highlight their importance in studying low cloud radiative forcing and low cloud climate feedback.