Abstract
A rain evaporation and downdraft parameterization is designed to complement the cumulus convection scheme of the Goddard Laboratory for Atmospheres General Circulation Model (GLA GCM). The scheme invokes (i) a diagnostic determination of the commencement level of rain-evaporation-induced downdrafts, (ii) a method for calculating downdraft mass fluxes emanating from different levels of the atmosphere, and (iii) an explicitly prescribed overall fraction of rain evaporation within the downdraft.
The parameterization was tested with the GATE [GARP (Global Atmospheric Research Program) Atlantic Tropical Experiment] phase III data in a fully prognostic mode and with the entire atmospheric and surface forcings prescribed with data. It was found that the near-surface downdraft cooling largely mitigates the observed surface sensible heating. In the absence of this cooling, the boundary layer must get rid of the surface heat flux by spurious turbulent transport, which becomes significant in simulations that ignore both the rain evaporation and downdrafts. The time mean as well as root-mean-square errors in the vertical temperature profiles are somewhat larger for simulations without downdrafts and are much larger for simulations without both downdrafts and rain evaporation. The downdrafts are found to produce a useful correction in the simulated near-surface temperature and humidity in GCM simulations, and in that way, the parameterization improves the simulation of tropospheric temperature and humidity. In a one-year comparison of GLA GCM simulations with and without downdrafts, the former produced better precipitation climatology and surface temperatures.
