Feature-based diagnosis of convective organization and cold pools using ARM data sets and evaluation of a unified convection parameterization

 
Poster PDF

Authors


Wei-Yi Cheng — University of Washington
Angela K Rowe — University of Wisconsin
Sungsu Park — Seoul National University

Category

Convective clouds, including aerosol interactions

Description

The two-way feedback between convective updrafts and cold pools has been suggested as a critical mechanism for the transition from shallow to deep convection over land and ocean, as well as for maintaining the deep convection phase. The unified convection scheme (UNICON, Park, 2014) is one cumulus parameterization scheme that explicitly represents this interaction between convective updrafts and cold pools. In UNICON, the degree of mesoscale convective organization is represented with a scalar variable and is proportional to the fractional area covered by cold pools, which is triggered when the convective saturated downdraft is strong enough to penetrate into the boundary layer. When the degree of organization is high, convective plumes have lower entrainment rates and higher perturbation temperature, specific humidity, and vertical velocity at the surface where the plumes initiate. In this study, ARM observations are used to constrain the two-way feedback processes between convection and cold pools simulated by UNICON. Radar-based diagnosis of convective organization and cold pools over the tropical ocean are conducted using observations from the ARM MJO Investigation Experiment (AMIE)/Dynamics of the MJO (DYNAMO) field campaigns. A modified Steiner et al. (1995) convective-stratiform partitioning algorithm is applied to the gridded radar data set and the results are used to derive spatial scalar metrics indicative of the degree of convective organization. The scalar metrics include the number and mean size of contiguous convective echoes and the mean distance between convective pixels. Cold-pool characteristics, such as maximum diameter, lifetime, and fractional coverage, are inferred during a limited period from radar observations. WRF simulations used in the Feng et al. (2015) study are analyzed in the same manner here to derive metrics of convective organization and cold-pool properties not available from observations. The AMIE-Gan forcing data set is used to force a single-column model (SCM) that is coupled with UNICON. Field variables from UNICON, especially those that represent the degree of convective organization and cold-pool properties, are compared against observations and WRF simulations in a qualitative manner. Possible causes of the discrepancy between observations and UNICON are discussed.