@misc{81345, author = {David Duncan and Niels Bormann and Alan Geer and Peter Weston}, title = {Superobbing and Finer Thinning for All-sky Humidity Sounder Assimilation}, abstract = {

Humidity sounder radiances are currently thinned to about 110 km spacing prior to assimilation at ECMWF and used at native resolution. In this paper, the thinning scale and possible averaging of all-sky humidity sounder observations are considered. Averaging the radiances into “superobs” acts as a low-pass filter and provides smoother images of departures. This decreases effective sensor noise and thus std(O-B), marginally for 183 GHz channels (5-15%) and significantly for 118 GHz channels (5-55%). The method of superobbing is argued to provide more representative observations for assimilation and a better utilisation of total information content than thinning native-resolution
radiances, as there is less information discarded prior to assimilation.


Thinning and averaging scales of humidity sounder radiances are investigated separately and then together. Superobbing of humidity sounder radiances and decreased thinning are shown to markedly improve background fits to independent humidity- and wind-sensitive observations, when changed in isolation or in combination. This is indicative of better short-range forecasts of humidity, with winds improved through the 4D-Var tracer effect. Results are relatively insensitive to superob resolution and so it is argued that smaller superobs are preferable for diagnostic purposes. The move from 110 to 70 km thinning provides as much as 140% more radiances for assimilation and much-improved spatial information for diagnostic purposes. These augmentations in the processing of humidity sounder radiances lead to slightly improved wind forecasts in the Southern Hemisphere at medium-range. As part of the upgrade to Cycle 49r1, all-sky humidity sounder radiances will be assimilated at 70 km spacing after being averaged into 50 km superobs.

}, year = {2023}, journal = {ECMWF Technical Memoranda}, number = {905}, month = {03/2023}, publisher = {ECMWF}, url = { }, doi = {10.21957/5c3b9c8d9f}, }