| description abstract | Heat and moisture budgets are computed for the 40 km model simulations of moist frontogenesis described recently by Hsie and others. The apparent heat source and moisture sinks are dominated by the condensation term and have maxima in the middle troposphere. Both the large-scale (200 km) moisture convergence and the large-scale vertical motion are highly correlated with the mesoscale condensation rate. Four alternative schemes for treating the effects of moist convection in primitive equation models are tested, and the results compared with those from the explicit scheme for calculating condensation and precipitation. A scheme in which only water vapor is predicted yields results similar to the explicit simulation, which included prediction equations for water vapor, cloud water and rainwater. The neglect of two opposing effects-water loading and evaporation-is apparently responsible for the similarity to the control. However, when both cloud water and water vapor are predicted, the presence of evaporation, but not water loading by rain, results in larger differences from the control. A third scheme, developed by Kessler, does not conserve total water, and the latent heat released is overestimated. The cumulus parameterization proposed by Anthes is tested as the fourth scheme in a coarse-resolution (200 km) version of the model. The major deficiency with the coarse-mesh model is its failure to resolve a narrow, frictionally-driven updraft close to the surface cold front. The feedback of the moisture convergence and the cumulus heating parameterization produces large differences in the horizontal distribution of heating in the thermal and wind structure. When the cumulus heating in the coarse-mesh simulation is specified from the average distribution given by the explicit scheme in the fine-mesh simulation, the coarse-mesh simulation is much closer to the large-scale average of the fine-mesh simulation. | |
