Development of a Quasi-Global Fundamental Climate Data Record for Observations from Geostationary Satellites

The utilisation of observations of past, present, and future geostationary satellites for climate monitoring is a challenge. Since the late 1970s, space agencies operated up to 50 geostationary satellite missions with a variety of instrumentation. Merging these observations in a quasi-global geostationary 'ring' data record is essential for the provision of satellite-based data records of Essential Climate Variables (ECVs). EUMETSAT is engaged in data rescue, uncertainty characterisation, recalibration, and harmonisation of these observations and aims at the provision of the data to users on its joint EUMETSAT-ECMWF cloud infrastructure the so called European Weather Cloud and the EUMETSAT Data Store. The process of preparing satellite data for climate monitoring and analysis - such as undertaken by WCRP’s project GEWEX - is tedious and only recently being recognised as fundamental first step in preparing records ECVs from these data. 

Past and present geostationary data come with the possibility of unforeseen radiometric, geometric, and metadata anomalies. These anomalies may be related to the instrument or the data processing. EUMETSAT developed a system that performs an automatic anomaly analysis to the observations of past and present Meteosat and JMA satellites. The system is able to detect the most common types of anomalies with a high probability of detection and low false alarm rate. The anomalies are stored in a data base so as to inform downstream processing. As the anomalies are flagged on a pixel-by-pixel basis the loss of data is kept to a minimum.

EUMETSAT recalibrated its anomaly screened infrared channel observations from MVIRI on Meteosat First Generation (MFG) and SEVIRI on Meteosat Second Generation (MSG) measurements against IASI, AIRS, and HIRS measurements. The recalibration improved the radiometric accuracy of MVIRI and SEVIR to less than 0.5 K. Such improvements allow the seamless use of these observations for the retrievals of ECVs data records from geostationary orbit covering more than 40 years. Similarly, EUMETSAT applied its recalibration approach to the instruments operated on JMA’s geostationary satellites, resulting in similar improvements as made for the Meteosat satellites. Regarding satellite data quality, first steps have been made to provide recalibrated data with quantitative uncertainty estimates, as developed in the framework of the EU-H2020 FIDUCEO project. Such estimates add another dimension of quality information that is essential to make a data record a true climate data record.  With the aim to close the geostationary 'ring', EUMETSAT and NOAA now started applying the methods presented above to the US geostationary sensor data as well.

Once available, the individual time-series of recalibrated geostationary satellite data of the three collaborating organisations (EUMETSAT, JMA, and NOAA) will be quality controlled, cross-calibrated and merged into a single geostationary ‘ring’ product. Hereto the methods developed by the ISCPP-NG will be used. The collaborating organisations plan to use the cloud computing infrastructure to work on the data that are distributed over three continents.