The Role of Negative Buoyancy in Surface-Based Convection and Its Representation in Cumulus Parameterization Schemes

Several warm season, late-afternoon precipitation events are simulated over the Chesapeake Bay watershed using the Weather Research and Forecasting (WRF) model at three different resolutions. The onset and peak of surface-based convection are predicted to occur prematurely when two popular cumulus parameterization schemes (Betts-Miller-Janji and Kain-Fritsch) are used. Rainfall predictions are significantly improved with explicit convection. The early bias appears to be associated with the inadequacy in representing convective inhibition (CIN) or negative buoyancy in the trigger for moist convection. In particular, both schemes have weak constraints for the negative buoyancy above cloud base and below the level of free convection, leading to premature rainfall. Satellite-derived soundings suggest that, even with extremely favorable conditions, negative buoyancy in this layer may delay the onset of surface-based convection. Other factors, such as enhanced mixing due to overactive shallow convection, also appear to contribute to the early rainfall bias through the premature removal of CIN during the day.