We are using observations, cloud-resolving model simulations, and single-column model simulations to better understand how to parameterize the interaction between deep cumulus convection and the boundary layer over the southern Great Plains of the United States. The observations are from a 29-day DOE ARM (Atmospheric Radiation Measurement program) Single Column Model (SCM) Intensive Observation Period (IOP) that took place at the ARM Southern Great Plains (SGP) site during June and July 1997.

The SCM IOP observations included temperature, humidity, and wind profiles from radiosondes launched at 3-hourly intervals from the Central Facility (CF) and the four Boundary Facilities, surface turbulent and radiative fluxes, rainfall rates based on a combination of radar and rain gauge measurements, top-of-atmosphere radiative fluxes, cloud amounts, and cloud fraction profiles obtained from a cloud radar. The surface turbulent fluxes were measured every thirty minutes by 10 Energy Balance Bowen Ratio stations located at the CF and several of the Extended Facilities.

The ARM Data and Science Integration Team processed the observations using a constrained variational analysis technique in order to obtain estimates of the advective tendencies of temperature and water vapor averaged over the SCM domain, an area that corresponds approximately to the SGP site. These estimates, along with those of the surface turbulent fluxes and the radiative heating rate profile, make it possible to perform diagnostic studies of the interaction between convection and the boundary layer, as well as simulations of this interaction using CRMs (cloud-resolving models) and SCMs. We have undertaken all three approaches.

The observations show that cumulus effects in the boundary layer are significant, and approximately balance those due to turbulence. In the CRM's boundary layer, drying by cumulus transport exceeds the moistening by rain evaporation, while both cumulus transport and rain evaporation cool near the surface. The effects of cumulus convection and turbulence in the boundary layer in the NCEP (National Centers for Environmental Prediction) SCM are in general agreement with those in the CRM. The surface fluxes of water vapor are much larger, while those of sensible heat are smaller, in the NCEP SCM than in the CRM. This leads to larger cumulus and turbulence effects on the water vapor budget (at all levels) and smaller effects on the sensible heat budget (near the surface and above 1500 m altitude) in the NCEP SCM than n the CRM.