Arctic Radiation and Turbulence Interaction STudy



ARTIST was an European Community project carried out by two Italian, two Finnish and two German partner institutions, subdivided into eight research groups. The core activity of the program was an extensive field study in the environment of the Spitzbergen Islands during March / April 1998.
The exchanges of momentum, energy and matter between the ocean and the atmosphere in the Arctic region have a significant impact on the regional and very likely also on the global atmospheric and oceanic circulations and consequently on the Earth’s climate. Therefore, the main concerns of ARTIST were detailed investigations of these exchange processes particularly at the partly ice-covered sea surface and their effects on the heat budget of the lower atmosphere.

In the framework of ARTIST our task was to provide a method to estimate sea surface radiative balance from satellite data in the Arctic region. Our approach to the problem was the utilisation of derived satellite parameters in bulk formulae. The determination of the short-wave and long-wave incoming radiation depends either on the right choice of the bulk formulae and on the correct value of the cloud cover. The latter was obtained with the aid of a cloud detection algorithm applied to AVHRR data. This cloud detection algorithm is probably one of the most important results of this project.


The algorithm (Ananasso et al. 2002) is a simplified version of Ebert’s (1987) method, it was tuned to typical early spring conditions in the ARTIST experimental area. It is able to discriminate between ice, cloud and open water situations. The algorithm has been applied to 154 scenes of NOAA14 satellite (from 16 March 1998 to 15 April 1998) covering the Arctic Ocean near Svalbard. The cloud detection results were validated by various independent procedures. According to a visual classification by an expert about 70% of the images (109) have an error less then 5% and only 11% of the images results have an error greater than 10%. The method was also tested against independent cloud and ice observations obtained respectively at the Ny-Ålesund base and from the SSM/I data set. The comparison with these independent sources of data confirms the good performance of the algorithm.


Several bulk formulae, typically used in Polar conditions, were tested by comparing the estimated insolation and atmospheric radiation fluxes with direct measurements made with radiometers mounted on board the Polar 2 aircraft and at the research base at Ny-Alesund. The comparison with direct observations permitted to evaluate the capability of NOAA/AVHRR satellite data to estimate the surface radiative fluxes at the air-sea interface. Moreover, this permitted us to define the best parameterizations to be applied in conjunction with satellite data to estimate both the short and long-wave downwelling radiation.


Differences between measurements and estimates can be attributed to cloud detection errors, natural variability of cloud conditions over the intercomparison areas but also on the simplicity of the parameterization schemes (Lind et al, 1984). Low bias of the mean value together with a low value of the RMSE in comparison to Key (1996) and other authors, indicates that the use of the cloud cover and surface temperature obtained from satellite measurements allow good estimates of the downwelling short and long-wave radiation of the Earth’s surface. This is especially true for time-averaged estimates in which the RMSE can only decrease. Obviously our detection method to estimate the cloud cover fraction seems to be superior to subjective ground based observations.


Finally a monthly averaged map of the total radiative balance of the sea surface has been mapped. In this case the resolution of the AVHRR data has been decreased to take into account the cloud cover. The balance is strongly influenced by the solar radiative term and by the sea surface temperature. The balance is positive everywhere with a minimum value in the northern part of the region (10-15 W/m2) and a maximum value along the Scandinavian cost (95 W/m2) where presumably the warm influence of the North Atlantic Current is still present.