Abstract
Three 15 h LAMPS (Limited Area Mesoscale Prediction System) model simulations were performed on a 70 km grid to examine the role of convective parameterization on the evolution of the Gulf Coast precipitation system of 6–7 March 1982. The first simulation neglected moist processes. The second used an explicit representation for moist physics as suggested by Orlanski and Ross (1984) and Rosenthal (1978; 1979). A third simulation employed the LAMPS sequential plume cumulus cloud model as convective parameterization for the mesoscale model. Results from these experiments indicate that the manner in which moist thermodynamics is included in a meso-alpha scale model has significant bearing on model depiction and evolution of mesoscale structures.
The dry simulation developed and maintained a low pressure area consistent with the observed Gulf Coast precipitation system; however, kinematic and thermodynamic fields were substantially modified by the inclusion of moist physics. The simulation with parameterized convection outperformed that with grid-resolved moist physics in several important respects. While precipitation in the grid-resolved simulation formed in the same location as the parameterized convection and as observed, it was slower to develop and was initiated about 2 h late over the Gulf Coast During the latter half of that simulation, the Gulf Coast system underwent explosive development, giving midtropospheric diabatic heating rates and vertical motions that were unrealistically large near the center of the main precipitation system. A superadiabatic layer was generated near the top of the model precipitation system for this experiment in response to the extreme midlevel heating. Overdevelopment appeared to be a result of a runaway CISK-like feedback between low-level moisture convergence, upward vertical motion and latent heat release. These difficulties reflected the inability of grid resolved explicit model dynamics to eliminate gravitational instability except through the generation of a grid-resolved convective circulation. The convective parameterization effectively acted as a convective adjustment to preclude development of the superadiabatic layer that led to overdevelopment in the explicit simulation.
