|Title||First year report: The impact of HIRS on ECMWF forecasts, adding ATMS data over land and sea ice and new observation errors for AMSU-A|
|Series/Collection||EUMETSAT/ECMWF Fellowship Programme Research Reports|
|Authors||Lawrence, H, Bormann, N|
|Event Series/Collection||EUMETSAT/ECMWF Fellowship Programme|
|Place of publication||Shinfield Park, Reading|
In this report we present three separate studies. In the first study we assessed the impact of the HIRS instrument on forecasts by performing assimilation trials. The HIRS instrument was found to have a small positive impact on short-range temperature and geopotential forecasts in the Southern Hemisphere and to improve the fit of MHS, AIRS and IASI observations to 12-hour forecasts. In the second study we extended the coverage of the ATMS instrument by adding surface-sensitive temperature and humidity sounding channels over land and sea-ice, following methods developed previously for MHS and AMSU-A. Results of adding data over land showed a mainly neutral impact on global forecast scores but improved the fits of 12-hour forecasts to MHS observations. As a result of this positive impact, ATMS surface-sensitive channels have been added over land operationally in the 40R3 ECMWF cycle. Results over sea-ice were more mixed, with an improvement in the fits of 12-hour forecasts to MHS observations but a slight degradation in day 1 and 2 temperature forecasts over the Southern Hemisphere. The third study presented in this report is the development and testing of situation-dependent observation errors for AMSU-A channels 5 to 7. In this study we developed new observation errors for these channels which account for forward model errors due to undetected cloud as well as errors in the estimated surface emissivity. As a result observation errors are higher over land and sea-ice, compared to over ocean, and in areas of high liquid water path and high surface-to-space transmittance. These new observation errors produced a neutral impact without changing the current screening but when the cloud screening was also relaxed forecast scores were improved by 1 2% for 500hPa Southern Hemisphere geopotential height and fits of 12-hour forecasts to ATMS observations were also improved by 0.1%. This last study is still ongoing.