In the framework of the GEOLAND project, EARS and CNRM/Météo-France assessed DSR from geo-stationary sensors data using different approaches, in order to match the requirements of Observatory Natural Carbon (ONC), which uses DSR to quantify the atmospheric forcing in the carbon cycle model, and of the Observatory Food Security and Crop Monitoring (OFM), which uses DSR in productivity modelling.

DSR product from METEOSAT images over Euro-Mediterranean region, first 10-day period of August 2004.

The EWBMS (Energy and Water Balance Monitoring System) database provided by EARS contains a DSR flux product derived from METEOSAT visible [0.3 - 1.05 µm] images at 0.04° space resolution. Since only noon images are used, a Fourier analysis of the daily solar cycle allows to relate the noon value to the daily average. For a clear pixel, a modified version of the Kondratyev model (1969), which takes into account the reflection and absorption in the atmosphere, relates the DSR flux at noon to the intensity of solar radiation at the edge of the atmosphere (1355 W.m-2) and to the solar radiation transmission through the cloud free atmosphere. To obtain a daily DSR flux product, the conversion factor is determined by integration of the daily solar cycle and is a function of latitude and day of year. When a pixel is cloudy, the solar radiation transmission through the clouds is calculated from the observed cloud albedo according to a relationship derived from the Kubelka-Munk theory. Here, the calculation of the DSR is similar to the case of cloud-free pixels, except that the atmospheric transmission factor is replaced by the cloud transmission factor. The DSR flux is calculated on a daily basis and then for longer time periods by averaging (see figure above). This DSR flux has been validated by comparison with field measurements.

To get the products provided by EARS, it is necessary to make a request at the address ears@ears.nl.

CNRM/Météo-France set-up an algorithm shared in two scenarii, one for clear pixels, and one for cloudy pixels. In the case of clear pixels, an innovative method based upon the generation of a Look-up-Table (LUT) was developed to simulate the up-welling transmitted radiation as well as the DSR in varying atmospheric compositions (water vapor, ozone, aerosols). This approach was tested over some test cases acquired by METEOSAT-7, and validated by compariosn with the ground measurements of the BSRN (Baseline Surface Radiation Network) network (Elias and Roujean, 2007). The LUT was adapted to the cloudy sky case, using a cloud mask provided by the University of Karlsruhe (IMK). However, because it is necessary to optimize the generation of the LUT to reduce the computing time for operational purposes, this method has not been applied at large scale during the project life.

Kondratyev, K. Y., Radiation in the atmosphere, New York, London : Academic Press, 1969.

Elias, T. and J.L. Roujean, Estimation of the aerosol radiative forcing at ground level, over land, and in cloudless atmosphere, from METEOSAT-7 observation : method and case study, accepted in Atmospheric Chemistry and Physics, 2007.