PERFORMANCE OF THE ATMOSPHERIC CORRECTION OF MERIS DATA OVER NORTHERN EUROPEAN WATERS

摘要

The Northern European countries are particularly interested in monitoring water quality due to its source for fishery and tourism and due to the threats of accidental pollution. Water quality parameters derived from optical measurements require an excellent atmospheric correction, which is particularly difficult for northern European waters (North Atlantik, Bothnian Bight). The problems are manifold: between November and February the sun is very low and often exceeds a sun zenith angle of 65° where the plan-parallel assumption, which is used in most algorithms, reaches its limits. During the summer months, the sun glint limits the usage of a large portion of the MERIS swath. Clouds often cover large parts of the area. The air is known to be clear; reducing the problem of aerosol correction, however, very few measurements of aerosol properties of Northern Europe exists for validation. MERIS FR data are of great value for Norway because of the importance of the Fjords for fish farming and tourism. However, close to the shore, the environment effect is becoming very important because of the fragment shape of the coast. An arbitrary point in mid-Norway is measured by one, and every third day by two MERIS overpasses per day, resulting in a total of 45 overflights during a cycle of 35 days. This situation is good compared with central Europe, but is more than balanced by the restrictions listed above. The performance of the atmospheric correction for MERIS is assessed by analysing RR and FR products of the Norwegian coast. The cloud screening, including the ice_haze flag for cloud borders, is found to be very severe leaving very little data for atmospheric correction. The aerosol optical depth and Angstrom coefficients are compared with Aeronet data. The results are not conclusive, but this could be due to the large distance to the closest Aeronet station in Norkoping. The retrieved water leaving radiance reflectances of open ocean look excellent. But approaching the coast the reflectance in bands 1 and 2 tend to be overcorrected. In a distance of 1-5km off the coast the reflectances are starting to be flagged invalid and are partly getting negative, probably because of the environment effect. Recommendations for improving the atmospheric correction over such water types are formulated.