Abstract
A one-dimensional time-dependent cumulus model is developed and discussed. Data predicted by the model along with a bulk entrainment model are compared with a case study observation and Warner's mean profile of Q/QA. While a great deal of the discrepancy between observed and predicted data can he attributed to the transient nature of convection, the consistent pattern of overprediction of such cloud properties as Q/QA and vertical velocity is indeed disturbing. It is concluded that neither the entrainment model nor the scalar nonlinear eddy viscosity model can adequately treat the general problem of turbulent transport in convective clouds. There is, however, sufficient evidence suggesting the models can he of practical value if their use is limited to dynamically active clouds and, in the case of the entrainment model, to a restricted portion of the cloud cycle life. Furthermore, there is little doubt that the entrainment coefficient is not a universal constant while the universality of the mixing length coefficients in the eddy viscosity models is still in question.
