|Title||All-sky assimilation: better snow-scattering radiative transfer and addition of SSMIS humidity sounding channels|
|Publication Type||Technical memorandum|
|Secondary Title||ECMWF Technical Memorandum|
The all-sky assimilation of microwave radiances has benefitted from removing the largest remaining systematic error in the observation operator. Previously, the optical properties of snow hydrometeors were computed using Mie theory but now it is possible to use the Discrete Dipole Approximation, which gives more accurate results for complex 3D particles like snowflakes. These improvements in the observation operator allow the all-sky assimilation to be extended into deep-convective areas and to the 183GHz water-vapour sounding channels on SSMIS (Special Sensor Microwave Imager Sounder). There are clear benefits to midlatitude wind and geopotential forecasts as well as to humidity throughout the troposphere. Improved dynamical forecasts presumably come about because the forecast model inside 4D-Var is forced into a better fit to the observations of water vapour, cloud and precipitation. Scores are particularly helped by the mid and upper-tropospheric 183GHz water vapour channels, likely through the wind tracing effect. The beneficial impact on winds cannot be replicated using a clear-sky assimilation strategy. This encourages efforts to widen the application of the all-sky approach to other microwave instruments with water-vapour, cloud and precipitation tracing capability such as MHS (Microwave Humidity Sounder). Further, it should encourage attempts to extend all-sky assimilation to the water vapour channels on infrared instruments.