NWP assessment of EPS-SG MWS at ECMWF

The first MicroWave Sounder (MWS) instrument of the EUMETSAT Polar System Second Generation (EPS-SG) programme is evaluated in this report, with its performance and assimilation impact assessed using the ECMWF data assimilation system. MWS is a European-built instrument that carries on the tradition of previous operational sounders like AMSU-A and ATMS, sampling the 50-60 GHz oxygen band for temperature sounding and the 183 GHz band for humidity sounding, complemented with a set of suitable window channels. The Metop-SG-A1 satellite that carries the first MWS unit was launched in August 2025 and will be renamed Metop-D upon its commissioning. The satellite provides important continuity of coverage in the 9:30 morning orbit with enhanced capabilities, including more channels and finer spatial resolution.

The calibration of MWS is in line with heritage instruments, as evaluated using the IFS short-range forecast coupled with RTTOV v14 for radiative transfer, i.e. analysing mean background departures. For the requirements evaluated here, the calibration accuracy generally meets end-user requirements, as specified in the EURD. Several of the temperature-sounding channels exhibit non-negligible biases as a function of orbital position; while these biases are within specification they nevertheless require correction before assimilation. Temperature and humidity sounding channels’ global biases are generally within }0.5 K of the IFS and within a few tenths of a Kelvin of other operational sounders, with a tendency for MWS to be slightly warmer. 

MWS temperature sounding channels perform well, with effective noise performance similar to ATMS for tropospheric sounding and significantly better for stratospheric sounding. The 183 GHz suite on MWS had excellent performance up until an anomaly occurred on-orbit in early March of 2026, after which gain was significantly reduced and radiometric noise increased to levels well outside of specifications and far worse than the performance of heritage instruments. This degraded noise performance of the 183 GHz channels greatly reduces their assimilation potential in clear skies, though the measurements in this band still retain some useable information in cloudy scenes that is accessible through all-sky assimilation.

Trials have been conducted with MWS assimilated in various configurations using the all-sky assimilation method. The report focusses on trials that cover the period after the 183 GHz anomaly in early March, with results from earlier trials prior to the anomaly provided in an appendix. Assimilation trials indicate a good forecast impact from MWS. Positive impacts on temperature and humidity are seen in short-range forecasts, judged by fits to independent observations. Medium-range impact is positive for geopotential height and winds, particularly in the Southern Hemisphere and in the lower troposphere. The results also suggest that there is some limited benefit from including the 183 GHz channels in the assimilation despite their high noise levels. 

The assimilation configuration detailed in the report paves the way for operational assimilation of MWS on Metop-D later in 2026, as the results presented are clearly positive. However, this configuration was purposefully conservative, focussing on assimilation of channels that are assimilated from equivalent instruments already in ECMWF operations. Relative to ATMS and AMSU, there are channels on MWS that exploit new parts of the spectrum, and these should be a focus for assimilation in future work. Full exploitation of the rich information provided by MWS is a worthy goal, and this report aims to be a reference document for the community in pursuing this.