TY - RPRT AU - Liam Steele AU - Niels Bormann AU - Alan Geer AU - Marcin Chrust AU - David Duncan AB - Microwave temperature and humidity sounding data are vital for improving numerical weather prediction skill, and are one of the most important observation types in the assimilation scheme used at ECMWF. In this report we present developments to the ECMWF Integrated Forecasting System (IFS) that allow for inter-channel error correlations and variational quality control to be used with all-sky observations from the Advanced Technology Microwave Sounder (ATMS), where cloudy or precipitating observations are not screened out. We discuss the creation of different error correlation matrices and assess their impact on the forecast.
We find that assimilating ATMS in all-sky conditions without accounting for inter-channel error correlations can lead to some improvements compared to clear-sky assimilation. Adding inter-channel error correlations results in some further slight improvements, but these are mainly limited to the southern hemisphere wind forecast. Using correlation matrices derived from background departure statistics gives a similar forecast performance as using matrices derived with the Desroziers diagnostic. The results have been achieved without specific reconditioning of the observation error covariance matrix, and introducing the error correlations does not lead to a significant increase in the number of iterations needed for 4D-Var to converge. Using an error correlation matrix derived from a reconditioned diagnosed covariance matrix has a neutral impact on forecast performance and the number of iterations needed.
The capabilities developed are highly relevant for future instruments with a larger number of cloud and precipitation-sensitive channels, such as EUMETSAT’s Micro-Wave Imager and Ice Cloud Imager
on the Metop-SG satellites, ESA’s Arctic Weather Satellite and the future EPS-Sterna microsatellite constellation, and Spire’s Hyperspectral Microwave Sounder. The developments could also potentially realise greater benefit from instruments that are currently assimilated, as well as with future all-sky infrared assimilation. BT - Eumetsat/ECMWF Fellowship Programme Research Report CY - Reading DA - 12/2025 DO - 10.21957/53801337ec M1 - RR68 M3 - Eumetsat/ECMWF Research Report N2 - Microwave temperature and humidity sounding data are vital for improving numerical weather prediction skill, and are one of the most important observation types in the assimilation scheme used at ECMWF. In this report we present developments to the ECMWF Integrated Forecasting System (IFS) that allow for inter-channel error correlations and variational quality control to be used with all-sky observations from the Advanced Technology Microwave Sounder (ATMS), where cloudy or precipitating observations are not screened out. We discuss the creation of different error correlation matrices and assess their impact on the forecast.
We find that assimilating ATMS in all-sky conditions without accounting for inter-channel error correlations can lead to some improvements compared to clear-sky assimilation. Adding inter-channel error correlations results in some further slight improvements, but these are mainly limited to the southern hemisphere wind forecast. Using correlation matrices derived from background departure statistics gives a similar forecast performance as using matrices derived with the Desroziers diagnostic. The results have been achieved without specific reconditioning of the observation error covariance matrix, and introducing the error correlations does not lead to a significant increase in the number of iterations needed for 4D-Var to converge. Using an error correlation matrix derived from a reconditioned diagnosed covariance matrix has a neutral impact on forecast performance and the number of iterations needed.
The capabilities developed are highly relevant for future instruments with a larger number of cloud and precipitation-sensitive channels, such as EUMETSAT’s Micro-Wave Imager and Ice Cloud Imager
on the Metop-SG satellites, ESA’s Arctic Weather Satellite and the future EPS-Sterna microsatellite constellation, and Spire’s Hyperspectral Microwave Sounder. The developments could also potentially realise greater benefit from instruments that are currently assimilated, as well as with future all-sky infrared assimilation. PB - ECMWF PP - Reading PY - 2025 T2 - Eumetsat/ECMWF Fellowship Programme Research Report TI - Enhancing the exploitation of all-sky microwave sensors at ECMWF using inter-channel error correlations UR -   SN - RR68 ER -