Satellite observations in the infrared and microwave parts of the spectrum have long been assimilated into forecasting systems to help estimate the best possible initial conditions for global weather predictions. Assimilating radiances in the visible part of the spectrum, on the other hand, continues to pose many challenges. The reason lies in the complex respective interactions of cloud and aerosol particles with radiation at those wavelengths as well as the complex characteristics of the surface as a reflector of visible light. These complications make it difficult to develop ‘observation operators’, which convert model values into satellite observation equivalents. However, progress towards assimilating visible radiances has recently been made in the context of the ARAS (Aerosol Radiance Assimilation Study) project funded by the European Space Agency (ESA).
Assimilating aerosol data
As part of ARAS, an observation operator based on the Oxford‐RAL Aerosol and Cloud (ORAC) satellite retrieval scheme has been developed and incorporated into ECMWF’s Integrated Forecasting System (IFS) with the help of the RAL (Retrieval of Aerosol and Cloud) group. This operator includes look‐up tables in which reflectances at the top of the atmosphere are stored as a function of aerosol optical properties such as optical depth, single scattering albedo and asymmetry parameter as well as satellite viewing geometry and the position of the sun. Optical depth is associated with the amount of radiation scattered or absorbed by aerosols, while the single scattering albedo essentially gives an indication of the part that is absorbed and the asymmetry parameter of the part that is scattered. Observations used in the ARAS project are the level 2 aerosol visible radiances (reflectances) from the MODIS instrument on board the Aqua and Terra satellites. This is the first time that this type of observation has been assimilated in ECMWF’s atmospheric 4D‐Var assimilation system. While assimilating such observations is still experimental, the results show great potential for future operational implementation in the atmospheric composition forecasts produced by the EU‐funded Copernicus Atmosphere Monitoring Service (CAMS) implemented by ECMWF.
A dust event coming from the Sahara desert on 1 March 2017 illustrates what difference the assimilation of visible radiances can make. The figure shows total aerosol optical depth (AOD) at 550 nm in the IFS analysis without any assimilated aerosol data and with the assimilation of MODIS visible reflectances, as well as satellite‐derived AODs from MODIS at the same wavelength. In this event, MODIS observed two plumes. The first one, crossing the Atlantic Ocean at around 10°N, is underestimated in the analysis.
The assimilation of radiances increases AOD in the analysis to a level comparable to the MODIS data. The second plume in the Gulf of Guinea is mostly missed in the analysis without any assimilated aerosol data. The assimilation of the radiances brings clear benefits in this case.
Outlook
ARAS is scheduled to finish in April 2020, but its outcomes will hopefully be useful to other applications. For example the new release of the RTTOV observation operator includes an extension to calculate radiances in the visible part of the spectrum for cloudy conditions based on the look‐up table approach. From a formal point of view, treating clouds or aerosols via this approach is very similar. This implies that many of the tools developed in ARAS for aerosol visible reflectance assimilation could be adapted for clouds, provided the appropriate look‐up tables are used. The use of visible radiances for cloud assimilation would be a major step forward as these data are currently not assimilated at all at any operational numerical weather prediction centre, even though they provide crucial information on the state of the atmosphere in cloudy conditions. More research is still needed, but the results from the aerosol assimilation achieved in ARAS could open the way towards a fuller exploitation of visible radiances to improve numerical weather prediction.
