Environmental and internal controls on Lagrangian transitions from closed cell mesoscale cellular convection over subtropical oceans

Classifications of mesoscale cellular convection (MCC) for marine boundary layer clouds are produced using a supervised neural network algorithm applied to MODIS daytime liquid water path data. The classifier, used in prior studies, distinguishes closed, open, and cellular but disorganized MCC. This work uses trajectories in four eastern subtropical ocean basins to compare meteorological variables and the structures of boundary layers for trajectories that begin as closed cells but evolve into either open cells or disorganized cells or remain closed cells over one afternoon-afternoon cycle. Results show contrasts between the trajectory sets: Trajectories for MCC that remain closed cells are more frequently observed nearer coasts, whereas trajectories that break into open and disorganized cells begin farther offshore. The frequency at which closed cells transition to open cells is seasonally invariant. The fraction of trajectories that stay as closed MCC varies throughout the year in opposition to those that break into disorganized cells, so that their annual cycles are 180 degrees out of phase. Trajectories remain as closed cell more frequently in austral spring and boreal summer when the trade inversion is stronger. The closed-disorganized MCC breakup is associated with weaker subsidence, a weaker inversion, a drier free troposphere, and enhanced nighttime boundary layer deepening, consistent with a warming-drying mechanism. The closed-open transition occurs in meteorological conditions similar to closed-closed trajectories. However, prior to the transition, the closed-open trajectories exhibit stronger surface winds and lower cloud droplet concentrations and rain more heavily overnight. Results suggest that multiple, independent mechanisms drive changes in cloud amount and morphology.