Simulations of trade wind cumuli under a strong inversion

The fifth intercomparison of the Global Water and Energy Experiment Cloud S ystem Studies Working Group 1 is used as a vehicle for better understanding the dynamics of trade wind cumuli capped by a strong inversion. The basis of the intercomparison is 10 simulations by 7 groups. These simulations are supplemented by many further sensitivity studies, including some with very refined grid meshes. The simulations help illustrate the turbulent dynamics of trade cumuli in s uch a regime. In many respects the dynamics are similar to those found in m any previous simulations of trade cumuli capped by weaker inversions. The p rincipal differences are the extent to which the cloud layer is quasi-stead y in the current simulations, evidence of weak countergradient momentum tra nsport within the cloud layer, and the development and influence of an inci pient stratiform cloud layer at the top of the cloud layer. Although many e lements of the turbulent structure (including the wind profiles, the evolut ion of cloud-base height, the statistics of the subcloud layer, and the nat ure of mixing in the lower and middle parts of the cloud layer) are robustl y predicted, the representation of the stratiform cloud amount by the diffe rent simulations is remarkably sensitive to a number of factors. Chief amon g these are differences between numerical algorithms. These sensitivities p ersist even among simulations on relatively refined grid meshes. Part of th is sensitivity is attributed to a physically realistic positive radiative f eedback, whereby a propensity toward higher cloud fractions in any given si mulation is amplified by longwave radiative cooling. The simulations also provide new insight into the dynamics of the transitio n layer at cloud base. In accord with observations, the simulations predict that this layer is most identifiable in terms of moisture variances and gr adients. The simulations help illustrate the highly variable (in both heigh t and thickness) nature of the transition layer, and we speculate that this variability helps regulate convection. Lastly the simulations are used to help evaluate simple models of trade win d boundary layers. In accord with previous studies, mass-flux models well r epresent the dynamics of the cloud layer, while mixing-length models well r epresent the subcloud layer. The development of the stratiform cloud layer is not, however, captured by the mass-flux models. The simulations indicate that future theoretical research needs to focus on interface rules, whereb y the cloud layer is coupled to the subcloud layer below and the free atmos phere above. Future observational studies of this regime would be of most b enefit if they could provide robust cloud statistics as a function of mean environmental conditions.