Using information about small-scale variability in water vapor in a convective parameterization
| Pincus, Robert | NOAA-CIRES Climate Diagnostics Center |
| Donner, Leo | Geophysical Fluid Dynamics Laboratory |
| Batstone, Crispian | NOAA-CIRES Climate Diagnostics Center |
Category: Modeling
Deep convection is the process that most strongly affects the transport of energy and moisture in the atmosphere, and it produces clouds which have enormous impacts on the radiation budget and produce large amounts of precipitation. Convection is exceptional in that it is typically initiated by the moistest and/or warmest parcels in the boundary layer and covers a very small portion of any given domain. Because convection is sensitive to both the initial parcel properties and the environment in which the plumes rise a wide range of convective clouds might arise even in a relatively small area. Parameterizations of convection in large-scale models, however, use information only from the mean thermodynamic sounding (since this is usually all that's available) and use an ensemble of plumes (typically constructed in somewhat arbitrary ways) to reproduce some aspects of the observed variability of convection. In this poster we explore ways in which a deep convective parameterization might be linked to an assumed-PDF scheme that predicts the large-scale distribution of total water (vapor +condensate) within each grid cell of a large-scale model. In the long run we plan to investigate ways in which knowledge of the variability in water vapor at each level might be used to improve predictions of triggering (the circumstances under which convection occurs) and closure (the determination of how much mass is transported by each plume). This poster address a simple question: can the range of behavior produced by an ensemble of plumes, each operating on the mean sounding but using a different entrainment rate, be reproduced by a single plume which sees the variability within some domain? We use as a testbed a month long cloud-resolving model simulation of summertime deep convection at the ARM SGP site. We compute temperatures, vertical velocities, mass fluxes, etc. using the plume model at the heart of the Donner (1993) parameterization of deep convection. We compare the range of plume properties from a) a set of plumes using a range of entrainment coefficients chosen to reproduce observed spectra of vertical velocity with b) the results from each plume rising in each non-convective column of the cloud resolving model. In most instances single, slowly-entraining plumes which experience a variable environment exhibit as wide a behavior as the ensemble of plumes acting on the mean sounding, suggesting the ensemble of entrainment coefficients may not be necessary of the subgrid-scale variability of water vapor is taken into account.
This poster will be displayed at the ARM Science Team Meeting.
