@manual{81347, author = {Tracy Scanlon and Alan Geer and Niels Bormann}, title = {Microwave Imagers in the ECMWF-IFS: Adding further observations and improving convective anvils in the observation operator}, abstract = {
Observations from microwave imagers provide a significant contribution to the European Centre for Medium Range Weather Forecasts Integrated Forecast System (ECMWF-IFS). Continued development of the framework that utilises these observations ensures that current and future instruments can be exploited to their full potential. In particular, the work in this study will benefit the upcoming EUMETSAT Polar System - Second Generation (EPS-SG) Microwave Imager (MWI) as well as the Ice Cloud Imager (ICI). This report focuses on the microwave imager data being assimilated (both in terms of instruments and frequencies) and the simulation of Brightness Temperatures (TBs) for comparison to the observations.
In this report, the use of data from previously unexploited channels of Special Sensor Microwave - Imager/Sounder (SSMIS) on board Defense Meteorological Satellite Program (DMSP) F18 (SSMIS-F18) is considered and it is found that there are some unexplained artefacts in the data relating to the use of the 37.0v GHz channel. These artefacts result in a 1.85 K bias between the ascending and descending nodes in the Southern Hemisphere winter. Excluding this channel and applying an Ascending / Descending (A/D) bias predictor reduces the adverse effects of including SSMIS-F18 on the 12 hour to 3 day forecast. Instead, the new channels give improved wind and humidity forecasts in the short range. Hence it is recommended that SSMIS-F18 be included into the ECMWF-IFS in this configuration.
The orbital biases seen in Microwave Radiation Imager (MWRI) on-board Fengyun-3D (FY-3D) are also discussed and it is demonstrated why this instrument has been removed from ECMWF operations from May 2022 onwards. A correction implemented by China Meteorological Administration (CMA) has resulted in a reduction but not elimination of the A/D biases, however, as these are still present the data will not be re-introduced into the operational system at this time.
This report also considers the inclusion of the 37h and 89h GHz-polarisation channels, which should bring new information on humidity and cloud. This is tested for currently assimilated microwave instruments. However, it is found that inclusion of the channels individually and together both results in degradation of the fit of the model to observations as well as the degradation of the forecast scores. This suggests that further development is required before these channels can be used, for example by improving observation error modelling or reducing systematic errors in the model. In terms of improving the simulated TBs, a new hydrometeor fraction assumption has been tested, which
models the convective precipitation fraction as linearly changing from the bottom to the top of the frozen part of an anvil cloud. This results in improvements in particular for the case where there were extreme differences between the observations and simulated TBs of up to ± 40 K; these improvements can be most clearly seen in the reduced skewness of the First Guess (FG) departures. Although these changes affect only a small number of observations, due to the large magnitude of the changes it will be implemented as part of a future version of the ECMWF-IFS.
Finally, the impact of extending the range of both the Liquid Water Content (LWC) and temperatures in the tables used for the optical properties in the simulation of TBs has been tested. This work is particuarly important to support the upcoming missions for EUMETSAT, including ICI. These have been shown to have a minimal impact on the simulated TBs, however in the future the extended temperature will become important, therefore, the functionality to allow this extension will be added to the RTTOV software package.