NL170_Digital_banner

Metop-A satellite retired after 15 years of huge benefit to forecasting

Mohamed Dahoui, Tony McNally, Niels Bormann, David Duncan

 

In November 2021, EUMETSAT decommissioned the Metop-A satellite, which has brought tremendous benefits to numerical weather prediction (NWP) since its launch about 15 years ago. Metop‑A was Europe’s first operational polar-orbiting satellite and the first in a series of three satellites of the EUMETSAT Polar System (EPS) programme covering the mid-morning orbit. It was packed with a suite of novel and advanced instruments, such as IASI, ASCAT, GRAS and GOME‑2, as well as valuable heritage sensors, such as AMSU‑A, MHS, HIRS and AVHRR. With the combined use of all these instruments, Metop‑A was arguably the most influential meteorological satellite of its generation. Follow-on Metop satellites (Metop‑B launched in 2012 and Metop‑C in 2018) offered extra benefits, useful resilience and new products based on the dual exploitation of satellite pairs. 

Improved forecasts

The combined use of data from the three Metop satellites represented nearly 45% of observations used at ECMWF and accounted for around 27% of the total impact of all observations assimilated by the Centre, as estimated by the Forecast Sensitivity to Observation Impact (FSOI) measure (see FSOI/data counts image). Having three Metop satellites provided additional impact and a healthy resilience in the observing system against instrument failure: it shielded the assimilation from intermittent outages affecting individual satellites or instruments. Two plots from 2013 show the big impact Metop‑A and Metop‑B had in terms of error reduction (see ‘Impact by hemisphere’ image).

FSOI and data counts in May 2020.
%3Cstrong%3EFSOI%20and%20data%20counts%20in%20May%202020.%3C/strong%3E%20Contribution%20of%20Metop%20instruments%20(shown%20in%20green),%20conventional%20observations%20and%20the%20rest%20of%20the%20observing%20system%20to%20the%20reduction%20of%20forecast%20error%20as%20estimated%20by%20FSOI%20(left),%20and%20used%20data%20counts%20(right),%20with%20error%20bars.
FSOI and data counts in May 2020. Contribution of Metop instruments (shown in green), conventional observations and the rest of the observing system to the reduction of forecast error as estimated by FSOI (left), and used data counts (right), with error bars.
Data from Metop helps to constrain many aspects of the atmosphere in the forecasting model: the vertical distribution of temperature, humidity and indirectly the wind (IASI, AMSU‑A, MHS and GRAS); surface winds (ASCAT); surface moisture over land (ASCAT); atmospheric winds (AVHRR); and ozone (GOME‑2). Derived products from IASI, GOME‑2 and AVHRR are used to monitor and constrain many aspects of atmospheric composition.

Metop instruments

IASI Infrared Atmospheric Sounding Interferometer AMSU-A  Advanced Microwave Sounding Unit–A
ASCAT Advanced Scatterometer MHS Microwave Humidity Sounder
GRAS GNSS (Global Navigation Satellite System) Receiver for Atmospheric Sounding HIRS High-resolution Infrared Radiation Sounder
GOME-2 Global Ozone Monitoring Experiment–2 AVHRR Advanced Very High Resolution Radiometer

Novel instruments

Aside from heritage instruments such as the microwave sounders AMSU‑A and MHS, Metop‑A introduced a range of new instruments which significantly extended observing capabilities.

The IASI instrument was the first infrared interferometer sounder with high spectral resolution (8461 channels). It was designed to measure temperature and water vapour profiles in addition to the concentration of other infrared-absorbing constituents. Throughout its lifetime, the instrument has proven to be fit for operations with stable performance and a consistent positive impact. Due to the configuration of assimilation systems and available resources, only an evolving subset of the most informative IASI channels was used. In addition to using radiances for NWP, ECMWF uses several datasets derived from IASI for producing the Copernicus Atmosphere Monitoring Service (CAMS) global analyses, forecasts and reanalyses, including CO, CO2 and CH4 and, in test mode, ozone and SO2.

ASCAT was a new-generation scatterometer instrument designed primarily to measure surface wind over oceans. The data are also useful to derive soil moisture over land. The instrument is equipped with two sets of three antennae allowing simultaneous measurements to be made from three directions to improve the wind direction ambiguity. Like IASI, ASCAT data started to be used not long after launch and has provided quasi uninterrupted service since then. For extended periods, ASCAT was the only instrument providing surface winds. Research is ongoing into the use of surface stress, which is closer to what the instrument measures but is not currently assimilated. Such work would allow ASCAT data to impact both the atmosphere and the ocean in the context of coupled data assimilation.

GRAS was the first operational European GNSS radio occultation instrument designed to measure temperature profiles and tropospheric humidity using refracted Global Positioning System signals. GRAS data are used by ECMWF in the form of bending angles and provide a consistent number of profiles. The consistent availability of GNSS-RO data helped fill large observation gaps in the stratosphere.

The GOME-2 instrument is designed to provide information on total column and vertical profiles of ozone as well as other species, such as NO2, SO2 and aerosols – the latter in combination with information also from IASI and AVHRR (Polar Multi-sensor Aerosol optical Properties product, PMAP). Total column ozone is used in the ECMWF atmospheric analysis system. More products (ozone, aerosols and SO2) are used by the CAMS analysis system.

Impact by hemisphere of adding two Metop satellites to the baseline system in 2013.
%3Cstrong%3EImpact%20by%20hemisphere%20of%20adding%20two%20Metop%20satellites%20to%20the%20baseline%20system%20in%202013.%3C/strong%3E%20The%20plots%20show%20the%20percentage%20reduction%20in%20root-mean-square%20error%20(RMSE)%20of%20500%20hPa%20geopotential%20height%20at%20days%201,%203%20and%205%20in%20the%20northern%20and%20southern%20hemisphere%20extratropics%20as%20a%20result%20of%20assimilating%20just%20Metop-A%20and%20both%20Metop-A%20and%20Metop-B.%20The%20error%20bars%20are%2095%25%20confidence%20intervals%20and%20represent%20the%20statistical%20significance%20of%20the%20improvements%20with%20respect%20to%20the%20%E2%80%98no%20Metop%E2%80%99%20baseline%20experiment.%20
Impact by hemisphere of adding two Metop satellites to the baseline system in 2013. The plots show the percentage reduction in root-mean-square error (RMSE) of 500 hPa geopotential height at days 1, 3 and 5 in the northern and southern hemisphere extratropics as a result of assimilating just Metop-A and both Metop-A and Metop-B. The error bars are 95% confidence intervals and represent the statistical significance of the improvements with respect to the ‘no Metop’ baseline experiment.

High reliability

EUMETSAT provided operational support for Metop satellites, leading to very high availability as well as speedy communication regarding anomalies and maintenance operations. EUMETSAT’s regional data service (EARS) and collaboration with the US National Oceanic and Atmospheric Administration (NOAA) to exploit the Antarctic dumping station at McMurdo allowed Metop data to be distributed to the user community in a timely manner, benefiting global and regional NWP models (see the AMSU‑A data figure).

The dual and subsequently triple exploitation of Metop satellites was performed in a way to maximise the benefits for NWP. This was achieved by choosing complementary orbits and generating dual products, such as global Atmospheric Motion Vectors. EUMETSAT managed to extend the lifetime of the Metop‑A satellite well beyond its life expectancy without noticeably compromising the performance of the instruments.

Metop‑A was a very successful operational satellite that had a tremendous impact on NWP. Indeed, its success has been key to securing substantial investment to launch the second generation of Metop platforms in the midtwenties. These will carry even more sophisticated technology to support operational NWP, but also to accelerate our understanding of Earth system science.

Mean timeliness of AMSU-A data from various satellites.
%3Cstrong%3EMean%20timeliness%20of%20AMSU-A%20data%20from%20various%20satellites.%3C/strong%3E%20The%20plot%20shows%20the%20mean%20timeliness%20of%20AMSU-A%20data%20from%20Metop%20compared%20to%20other%20satellites.
Mean timeliness of AMSU-A data from various satellites. The plot shows the mean timeliness of AMSU-A data from Metop compared to other satellites.