|Title||Assimilation of ATOVS radiances at ECMWF: second year EUMETSAT fellowship report|
|Series/Collection||EUMETSAT/ECMWF Fellowship Programme Research Reports|
|Authors||di Tomaso, E, Bormann, N|
|Event Series/Collection||EUMETSAT/ECMWF Fellowship Programme|
|Place of publication||Shinfield Park, Reading|
This report summarises work towards enhancing the use of satellite observations from microwave sounders in the ECMWF system over surfaces which are particularly difficult to model, such as land, high orography and sea ice. We have investigated the assimilation over land of MHS channel 5 radiances. Channel 5 is sensitive to low-level humidity as it is the lowest-peaking of the MHS channels in the 183 GHz water vapour band. Assimilation experiments show that the assimilation of channel 5 has a systematic impact on the mean moisture analysis which is consistent with independent GPS observations. In the second part of the report a new screening criterion has been tested to screen out radiance observations with too large a contribution from the surface. The new screening criterion is based on the transmittance from the surface to space and replaces the screening for high orography currently used for AMSU-A and MHS surface-sensitive channels, which rejects observations irrespectively of the instrument scan angle. Results of assimilation experiments suggest that a unique transmittance threshold per channel might not work globally. We suggest as alternative to the current screening, the use of an observation error which would take into account surface characteristics such as skin temperature and emissivity errors. In the last part of the report, we investigate an extension of the use of MHS data over sea ice. MHS observations are assimilated operationally only over surfaces with skin temperature greater than 278 K. As for the assimilation of microwave sounder data over land, using dynamically retrieved emissivities over sea ice provides a better representation of model-estimated brightness temperatures than a static scheme. We show that dynamic emissivities retrieved at 157 GHz lead to improved simulations for the 183 GHz sounding channels over sea ice allowing the assimilation of a considerable amount of observations in data sparse areas of the Globe.