EUMETSAT Fellows help improve ECMWF weather forecasts

1 April 2019

EUMETSAT Research Fellows hosted at ECMWF (as of March 2019)

EUMETSAT Research Fellows hosted at ECMWF investigate ways to improve the use of satellite data, including data from EUMETSAT’s Meteosat Second Generation satellites, a 1:4 scale model of which is shown in the picture. L-R: Katie Lean, Chris Burrows, Katrin Lonitz, Peter Weston.

EUMETSAT Research Fellows hosted at ECMWF joined Fellows based at other organisations at the annual ‘Fellow Day’ on 5 March 2019. The event at EUMETSAT’s headquarters in Darmstadt, Germany, was an opportunity for the young scientists to present their work on satellite data and share experiences from the past year.

EUMETSAT’s well-established Fellowship programme places early-career scientists at host institutions across Europe to develop new uses and applications of satellite mission products. Hosting Research Fellows through this scheme has been an important element of ECMWF’s close partnership with EUMETSAT for over 20 years.

Every day ECMWF uses tens of millions of satellite observations from space agencies across the world, including EUMETSAT. New instruments and better ways of interpreting the data continue to pave the way for better forecasts by improving the accuracy of the initial conditions that weather forecasts start from.

Four EUMETSAT Fellows are currently hosted at ECMWF. Their work spans research and operations: it ranges from assessing the quality of new satellite data before operational use to maintaining and improving the operational data assimilation system for the best use of such satellite data.

The next upgrade of ECMWF’s Integrated Forecasting System (IFS) planned for June this year will include several enhancements developed in EUMETSAT fellowship projects. The benefits are mutual: EUMETSAT reports that a survey conducted in 2017 showed 58 out of the 59 Fellows who had completed their EUMETSAT Fellowship since 1990 had found their way into the job market, including 48 in meteorology or atmospheric sciences.

Microwave imager radiances

Katrin Lonitz, who began her EUMETSAT Fellowship at ECMWF in 2014, works on the all-sky assimilation of microwave imager radiances, which provide information on a range of atmospheric, lands surface and ocean variables. ‘All-sky’ refers to assimilation under clear, cloudy and rainy conditions.

Katrin’s focus is on maintaining the operational assimilation of microwave imager data; extending their use in the IFS; improving formulations within the observation operator, which transforms model values into observation-like values; and the analysis of systematic biases related to clouds and precipitation.

Katrin contributed to the identification of a long-standing systematic model bias seen in extra-tropical storm tracks, where a lack of supercooled liquid water clouds in convective cold-air outbreaks caused model errors in simulated microwave radiance and shortwave radiation. She identified that the modelling of the electric properties of liquid water, especially supercooled water, needed an update.

The update has been incorporated into the latest version of the radiative transfer model (RTTOV-SCAT), which is maintained by EUMETSAT’s NWP SAF and is used to determine what kind of radiances would be measured given a particular state of the atmosphere. This means all users of RTTOV-SCAT will benefit from the improved modelling capabilities. It will also be implemented in the next upgrade of the IFS planned for June this year.

“In my remaining time as a EUMETSAT Fellow, I will tackle the assimilation of microwave imager radiances over land, because at present uncertainties in the land emissivity mean these data can be assimilated only over the ocean,” Katrin says.

Wind information from atmospheric motion vectors

Katie Lean started her Fellowship in October 2015, working on improving the interpretation and assimilation of satellite wind information from atmospheric motion vectors (AMVs).

AMVs form a valuable source of wind observations with five geostationary and eight polar-orbiting satellites currently providing good global coverage in ECMWF’s operational system.

While a long-established product, improved positive impacts from the data continue to be sought through gaining a better understanding of challenging areas and exploiting advances in satellite technology.

Typical coverage of active atmospheric motion vector (AMV) data for a 12-hour assimilation cycle (00 UTC, 7 March 2019)

Typical coverage of active atmospheric motion vector (AMV) data for a 12-hour assimilation cycle (00 UTC, 7 March 2019).

It has been an active period for new satellites, leading to a significant focus on the assessment and implementation of new AMV data in ECMWF’s operational system. The most recent addition has been the first of the newest generation of US Geostationary Operational Environmental Satellites (GOES), GOES-16.

“Ongoing activities to improve the use of AMVs in the current assimilation system particularly concern how to better account for errors in the assigned heights of the winds,” says Katie. “The advent of the Aeolus satellite brings an exciting opportunity this year to further explore features in the AMVs with a new source of tropospheric wind information.”

Infrared radiances

When new satellite observations become available, it is necessary to perform careful assessments of their quality. Offline assimilation experiments are performed prior to their operational assimilation, such as has been done for GOES-16 and Meteosat-11, whose data have become available recently.

Since 2017, EUMETSAT Fellow Chris Burrows has been working to improve the assimilation of infrared radiance observations from geostationary satellites, mainly Meteosat Second Generation, Himawari and the GOES series. Such observations provide information on atmospheric temperature and water vapour.

“I have prepared the ECMWF system to make better use of the geostationary satellite radiance data by introducing inter-channel error correlations into the assimilation system, activating slant-path radiative transfer and assimilating significantly more observations at high latitudes,” says Chris. These changes will become active in the IFS upgrade planned for June 2019.

More recently, ECMWF has been receiving – through the EUMETCast dissemination system – data from the GIIRS instrument on the Chinese satellite FY-4A, which is the first hyperspectral infrared sounder in a geostationary orbit.

Chris is working on a preliminary assessment of these observations: “We hope that the findings will be beneficial ahead of the launch of EUMETSAT’s Infrared Sounder (IRS) instrument on board Meteosat Third Generation, EUMETSAT’s next generation of geostationary satellites, set to launch from 2021 onwards.”

Meteosat Second Generation satellite in orbit

Geostationary satellites, such as this Meteosat Second Generation satellite, are in orbit nearly 36,000 km above the Earth’s equator. (Copyright: 2012 EUMETSAT)

Microwave sounding data

EUMETSAT Fellow Peter Weston works on improving the assimilation of microwave (MW) sounding data from polar-orbiting satellites, which provide information about atmospheric temperature and humidity.

Microwave sounding data from fourteen different satellites are currently assimilated into ECMWF's operational system. Peter is responsible for monitoring the data quality, and responding quickly to anomalies is crucial in maintaining the quality of the operational forecasts.

During his Fellowship, Peter has worked on various enhancements for the operational system, including harmonising the usage of different MW channels over snow, sea-ice, land and coasts and introducing the assimilation of Advanced Technology Microwave Sounder (ATMS) data from the NOAA-20 satellite launched in November 2017.

Work to take account of correlated observation errors for ATMS on board the Suomi-NPP and NOAA-20 satellites has resulted in 0.5–1% improvements in short- to medium-range forecasts of wind and geopotential height. This development will also be part of the IFS upgrade planned for June 2019.

“I have also investigated the assimilation of AMSU-A observations in the presence of cloud and precipitation,” Peter says. “There is still more work to do for this development to become operational, but significant progress has been made and my work has laid the foundations for future microwave temperature sounding instruments, such as the MWS on MetOp Second Generation, to be assimilated in all-sky conditions.”

Critical links

With their high level of expertise in satellite products and operational weather prediction, EUMETSAT Research Fellows are a critical link between EUMETSAT and ECMWF, helping to ensure that Member States and society gain maximum benefit from European investments in space observations.

Fellowship programme research reports from current and previous EUMETSAT Research Fellows hosted at ECMWF are available in our Publications library.