2002 Annual

Tuesday, 15 January 2002: 5:15 PM
Middle atmosphere data assimilation with a climate model
Saroja Polavarapu, MSC, Downsview, ON, Canada; and S. Ren, Y. Rochon, and D. Sankey
Poster PDF (114.2 kB)
A new data assimilation scheme is being adapted for the Canadian Middle Atmosphere Model (CMAM) for two reasons: (1) to confront the climate model with data in order to improve the model and reduce its errors, which will in turn lead to a better understanding of atmospheric dynamics and chemistry, and (2) to create a platform to aid in the testing and assimilation of data from new satellite instruments targeting the middle atmosphere (10-90 km). CMAM is a spectral T47L65 model with interactive chemistry, dynamics and radiation which produces a realistic ozone climatology and can produce QBO-like oscillations. The 3DVAR scheme which is used for operational weather forecasts at the Canadian Meteorological Centre has been adapted and coupled to CMAM. Necessary changes to the 3DVAR include implementing a generalized vertical coordinate, raising the lid to 0.001 mb from 10 mb, and introducing new background error statistics. As a first step, preliminary background error statistics are derived for each month from CMAM model climatological fields. The new assimilation scheme is being tested using past data from Jan 1994, during the UARS period, because of the availability of middle atmosphere data (e.g. from MLS, HRDI and WINDII) and because the climate of this period has been well documented. Because chemistry can influence CMAM’s radiation and dynamics, species assimilation is also planned for the UARS period with TOVS, TOMS and MLS as possible data sources. Results of the 3DVAR assimilations with a lid at 0.001 mb will be presented, along with plans for improvement of the system.

Supplementary URL: