Saroja Polavarapu, Shuzhan Ren, Yves Rochon, David Sankey
MSC and U of Toronto
CMC's operational 3DVAR scheme has been coupled to the Canadian Middle
Atmosphere Model (CMAM). CMAM is a comprehensive GCM with fully interactive
chemistry, radiation and dynamics extending from the ground to the top of
the mesosphere (95 km). During the past year, the system has continued to
evolve and we now employ the currently operational 3DVAR version in which
temperature and surface pressure are directly assimilated and which can
handle more types of observations. In addition, our period of study has
been changed from the UARS (Upper Atmosphere Research Satellite) period of
1991-4 to the current year (2002) in order to take advantage of the recent
and coming satellite data (AMSU, SSMIS, AIRS, ENVISAT, OSIRIS, EOS-AURA,
etc.). This configuration of the CMAM data assimilation system has been
validated against observations and UKMO analyses and is found to be a
distinct improvement over the previous version. This new configuration
also carries 15 species. Although no chemical data is assimilated, the
model updates the species distributions during the 6-h forecast to obtain
trial fields. Comparison of the update ozone field against observations
reveals a primarily zonal structure to the error. Nevertheless, the updated
ozone is clearly superior to those obtained from climate simulations.
Current work involves further diagnosis of the impact of dynamic variable
assimilation on ozone fields and identifying the source(s) of the zonal
error structure.