"Improving the representation of clouds in general circulation model (GCM)
simulations through analysis of cloud resolving model (CRM) results and 
field data"

Catherine Rio

The purpose of this study is to improve our understanding of the physical
processes involved in the redistribution of heat and moisture by tropical
convection, in order to adequately represent them in parameterizations of
deep convection for GCMs.
A now common way of evaluating parameterizations is to run two types of
simulations with the exact same large-scale forcings derived from
observational field campaigns: GCMs in a single column mode on the one
hand, and high resolution simulations over a domain equivalent to the GCM
grid size on the other hand. In the former, processes related to
convection and clouds are sub-grid scale and have to be parameterized,
while in the latter, the resolution is such that the cloud circulation at
convective scale is explicitly resolved. In order to improve
parameterizations, it is not sufficient to just compare the mean profiles
of temperature and humidity resulting from the two types of simulations,
it is necessary to analyze CRM simulations further, to extract
parameterizations internal variables, like for example properties of
updrafts and downdrafts carrying the vertical transport of mass, heat,
moisture, momentum...
Here, CRM simulations of tropical convective systems observed of the coast
of Australia during the monsoon season in 2006 (TWP-ICE campaign) are
analyzed in order to extract characteristics of downdrafts that are hardly
measurable but helpful to understand the physical processes related to
downdraft initiation and development as well as their role on the diurnal
cycle of deep convection. Those results are then used to evaluate the
parameterization of downdrafts in GISS modelE and propose further
development.