THE EVALUATION OF ENVISAT DATA USING A NUMERICAL WEATHER PREDICTION DATA ASSIMILATION SYSTEM

This paper discusses the use of the powerful technique of data assimilation to evaluate temperature, ozone and water vapour from the Envisat instruments: GOMOS, MIPAS and SCIAMACHY. The objective is to assess the error characteristics and biases of the Envisat data. Data assimilation will be used to combine the Envisat data with other contemporaneous measurements, and with data from an atmospheric general circulation model. To test the error characteristics and biases of the Envisat data, a number of procedures will be implemented: (1) Observing System Experiments (OSEs) in which different sets of data are systematically removed from the assimilation; (2) use of the statistics generated by the assimilation system to identify potential biases in the Envisat datasets; (3) statistical tests to assess the deviations from normality of the errors in the Envisat data and the assimilated datasets; (4) statistical tests to assess the characteristics of the residual (Envisat observations minus assimilated datasets); (5) independent comparison against data not used in the assimilation. To prepare for the receipt of Envisat data, the assimilation technique will tested with data from NASA's UARS and from ESA's GOME, and with synthetic data. Illustrative results from the assimilation of temperature, ozone and water vapour from the UARS MLS and the assimilation of total ozone from GOME will be presented.