Simulations of Observed Shallow-Deep Convection Transitions at the SGP

Description

The diurnal cycle of continental convection remains a challenge for most global models, which predict a transition from shallow to deep convection too early in the day. We have identified several promising shallow-deep convection transition cases that occur in situ, and thus have late afternoon maxima, in the SGP dataset compiled by Zhang and Klein (2010). For two of these cases we have simulated the transition with the WRF model run at 500 m resolution over a 250 x 250 km domain. In both cases the WRF simulates shallow convection at noon, a transition to isolated deep convection at 3:00 PM, and another transition to organized mesoscale convection by 6:00 PM. The timing is consistent with GOES images and ARSCL cloud profiles. Analysis of domain mean moist static energy profiles in one case indicates that the minimum distance needed to lift a parcel for it to become buoyant (~500 m) occurs near 2:00 PM, consistent with the simulated onset of deep convection. This case is thus a good target for the evaluation of the transition in Single Column Models. On the other hand, when shallow convection is developing near noon, the minimum distance is ~1.4 km, suggesting that either fairly vigorous boundary layer turbulence explains the existence of shallow clouds at this time or that they rely on subgrid variance of thermodynamic properties for their existence. We are exploring the ability of the GISS SCM to simulate these transition cases, using both large-scale forcing data in isolation and with relaxation to the instantaneous domain mean thermodynamic structure simulated by the WRF to understand whether the SCM can produce the correct transition, and if not, whether errors in the atmospheric state or inadequate parameterized moist convection physics is the source of the problem.