J5.4 The role of negative buoyancy in surface-based convection and its representation in cumulus parameterization schemes

Thursday, 10 January 2013: 9:15 AM
Room 10B (Austin Convention Center)
Manisha Ganeshan, University of Maryland, College Park, MD; and R. Murtugudde
Manuscript (8.4 MB)

Previous climate modeling experiments have reported significant biases in the representation of diurnal cycle of warm season precipitation. This study investigates potential reasons for some of these biases by using a combination of model sensitivity experiments and satellite observations in a regional modeling framework. Several warm-season, late-afternoon precipitation events are simulated over the Chesapeake Bay watershed using WRF at three different resolutions. The onset and peak of surface-based convection is predicted to occur prematurely when two popular cumulus parameterization schemes (Betts-Miller-Janjić and Kain-Fritsch) are used. Rainfall predictions are significantly improved with explicit convection. The early bias in parameterized convection appears to be associated with the deficiencies in representing convective inhibition (CIN) or negative buoyancy in the trigger for deep convection. Satellite-derived soundings suggest that even with extremely favorable conditions, the negative buoyancy in the layer above the cloud-base and below the level of free convection can delay the onset of surface-based convection. The trigger functions in both schemes fail to adequately account for this negative buoyancy, leading to premature rainfall. Other features that may also contribute to this early bias include an overactive mixing and a rapid growth of the convective boundary layer. The results from this study suggest that an enhanced representation of negative buoyancy in cumulus parameterization schemes may improve the simulation of the diurnal cycle of warm season precipitation.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner