| description abstract | Observational studies indicate that a mass flux approach may provide a realistic framework for cumulus parameterization in large-scale models, but this approach, through the introduction of a spectral cloud ensemble, leads normally to rather complex schemes. In this paper the question is addressed whether much simpler schemes can already provide realistic values of the thermal forcing by convection under various synoptic conditions. This is done through verifying such a scheme first on data from field experiments for periods of tropical penetrative convection (GATE, Marshall Islands), tradewind cumuli (ATEX, BOMEX) and extratropical organized convection (SESAME-79) and then in a NWP model. The scheme considers a population of clouds where the cloud ensemble is described by a one-dimensional bulk model as earlier applied by Yanai et al. in a diagnostic study of tropical convection. Cumulus scale downdrafts are included. Various types of convection are represented, i.e., penetrative convection in connection with large-scale convergent flow, shallow convection in suppressed conditions like tradewind cumuli and midlevel convection like extratropical organized convection associated with potentially unstable air above the boundary layer and large-scale ascent. The closure assumptions for determining the bulk cloud mass flux are: penetrative convection and midlevel convection are maintained by large-scale moisture convergence and shallow convection by supply of moisture due to surface evaporation. The parameterization produces realistic fields of convective heating and appears to be in fair balance with real data for NWP as it does not initiate strong adjustment processes (spinup) in global form. | |
