A comparison of explicit and implicit predictions of convective and stratiform precipitating weather systems with a meso‐β‐scale numerical model

Abstract The feasibility of the implicit (parametrized) compared to the explicit (grid‐resolved) approach for meso‐β‐scale models is examined using two different mesoscale precipitating weather systems (MPWSs): a squall line and a mesoscale convective complex (MCC) with an embedded meso‐vortex, that were responsible for the 1977 Johnstown, Pennsylvania, flood events. The prognostic explicit scheme contains predictive equations for cloud water and rainwater incorporating liquid water evaporation and hydrostatic water loading, and it is tested with grid sizes of 25 and 12‐5 km, respectively. Due to a near‐saturated and moist adiabatic initial environment associated with the MPWSs, there is no notable delay in the initiation of the grid‐box saturation. It is found that the explicit convective scheme fails to reproduce the convective precipitation related to the squall line and MCC, while it overpredicts the stratiform rainfall associated with the meso‐vortex. On the other hand, the implicit convective scheme reproduces the size, orientation and evolution of the squall line, but totally misses the meso‐vortex. A semi‐implicit control simulation (implicit plus diagnostic explicit) reproduces very well both the squall line and MCC. The results indicate that even with a grid resolution on the order of 10 km, an implicit convective scheme is very important for the numerical prediction of convectively driven precipitating systems. The roles of different model physics in controlling the frequent model ‘blow‐ups’ are also examined. It is found that if the semi‐implicit or the explicit convective scheme is used alone, the model sustains different degrees of overdevelopment of the mesocyclogenesis due to the CISK‐like feedback process among latent heat release, larger‐scale moisture convergence and the surface pressure fall. Model simulation with the diagnostic explicit scheme is the most affected by this instability, followed by the prognostic explicit scheme. The parametrized convection tends to reduce or eliminate this problem by elevating the heating and moistening maximum and removing moisture that otherwise would be used for explicit condensation. The liquid water evaporation is very effective in an unsaturated environment whereas the hydrostatic water loading has a significant retarding effect under saturated conditions. Intercomparisons of six experimental simulations in addition to the control run indicate that the combination of the implicit and prognostic explicit schemes (full physics) is superior in reproducing the basic sequence of convective and stratiform rainfall and the extent of the mesocyclogenesis associated with the Johnstown MPWSs. Because of their different roles in handling various types and scales of precipitation, the full physics approach appears to allow a broader scale interaction between the parametrized convection and mesoscale environment and to have the greatest potential success in predicting MPWSs that occur under various environmental conditions.