C3S data-Rescue


Multi-decadal climate data records (CDRs) are needed to improve our understanding of recent and future climate evolution. Constructing these records requires high quality and consistently processed observations such as, for example, global upper air data provided by satellite-based instruments. Observations from dedicated meteorological satellite missions launched over the last four decades are assimilated using state-of-the-art NWP models, specifically the data assimilation components of those systems, to generate high quality reanalyses such as ECMWF’s ERA5 (Hersbach et al. 2020). To extend the length of the CDRs further into the past, several recent projects have explored the potential of including measurements collected by earlier, pre-1979 satellite research instruments. 

Satellite data rescue is a recent activity that follows from the realisation that: (i) some early pre-1979 satellite datasets are being lost as they are stored on ageing media which can no longer be easily read; (ii) the collective knowledge of these old satellite missions is fading as the mission scientists retire and the documentation is being lost, and (iii) often important work is needed to prepare the data to a state where they can be assimilated.

A review prepared during EU-FP7 ERA-CLIM2 project identified early satellite datasets with high potential for assimilation in a future atmospheric reanalysis (Saunders et al, 2017). The factors to consider are: access to data record; length and continuity of record, quality (bias, variance, stability), overlap with other instruments, availability of observation operators to enable assimilation, sufficient metadata to be able to read and process the data, and the ability to screen the data for clouds and precipitation.

The recognition of the high importance of historical satellite data records for long-term reanalysis led EU’s Copernicus Climate Change Service (C3S) to launch its own Satellite Data Rescue projects with the objective to recover, assess and bring to a high level of preparedness a selection of early satellite data records for use in climate reanalysis, and in particular the next reanalysis ERA6 (scheduled for launch in 2024). For the identified early, high-potential data sets, the work done included performing quality checks, developing historical data readers and storing the data in a modern format (netCDF), carrying out various corrective steps such as adding quality flags, redoing navigation, recalibration, bias reduction, etc. Where needed, the radiative transfer tools were put in place to enable the simulation of radiances based on model background fields (e.g. definition of the Spectral Response Functions, calculation of the RTTOV coefficients). A dedicated task was also set up to generate an inventory of historical satellite data records to be used as a resource for coordination of current and future data rescue activities. The inventory contains standardized information on the pre-1979 sensors with interest for assimilation in climate models: a brief description of the sensor, mission objective, sampling and FOV resolution, channel characteristics, vertical sensitivities of channels, period of data availability,…

The objective of the web interface is to facilitate the access and use of the inventory by users interested in applications of early historical satellite data records for reanalysis as well as climate modelling and monitoring.

Hersbach, H, Bell, B, Berrisford, P, et al. The ERA5 global reanalysis. 2020, Vol. 146, pp. 1999-2049.

Saunders, R., P. Rayer and P. Poli. Update on Satellite Datasets for ERA-CLIM2. U.K. Met Office and Météo-France. 2017. p. 10. Version 1.1 29/01/17. Ref. Ares(2017) 802703 – 14/02/2017.