Data Assimilation with the Canadian Middle Atmosphere Model

A three-dimensional variational (3D-var) data assimilation scheme has been developed for the Canadian Middle Atmosphere Model (CMAM) in order to provide further insight into the model's errors and to provide a platform for the design and testing of satellite instruments targeting the middle atmosphere. CMAM is a 65-level (90 km) General Circulation Model of the troposphere-stratosphere-mesosphere capable of producing a realistic ozone climatology and QBO-like oscillations. Because CMAM has fully interactive chemistry, radiation and dynamics, the assimilation of ozone and other species is planned and may help to diagnose the interplay between these three processes. The 3D-var scheme has been adapted from the Canadian Meteorological Centre's operational weather forecasting system primarily by raising the lid, changing the vertical coordinate, adding analysis variables and developing new statistics. A preliminary set of background-error statistics have been estimated from CMAM climatology and are compared to statistics used by the operational Canadian weather forecast model. Results of the initial validation of the CMAM/3D-var system involving conventional meteorological observations from the troposphere and lower stratosphere, supplemented by middle atmosphere ozone and temperature data from satellite instruments such as TOMS and MLS, are also presented. Future plans include the assimilation of other chemicals and other sources of satellite data.