The assessment of optimal MERIS ocean colour products in the shelf waters of the KwaZulu-Natal Bight, South Africa

Abstract

The KwaZulu-Natal Bight is a highly variable bio-optical environment, where waters over the shelf can change from the oligotrophic case 1 conditions of the Agulhas Current to the case 2 inshore environment influenced by upwelling and riverine influx. This study represents the first radiometric and biogeochemical validation to be performed in the KwaZulu-Natal Bight. The aim is to assess the performance of the Medium Resolution Imaging Spectrometer (MERIS) normalised water-leaving reflectance ( (subw)), aerosol and chlorophyll a (Chl-a) products from the 2nd and 3rd reprocessing as well as the case 2 Regional (C2R) processor. Confidence flags indicated that the ocean colour products from the 2nd reprocessing were not reliable over the sampling site during the study period. Standard MERIS (subw) products from the 3rd reprocessing gave good returns from 490 to 560 nm, with absolute percentage difference (APD) of 10–16%, whilst underestimations in the red ranged from 124 to 215% compared to in situ data. Adjacency correction with the improved contrast between ocean and land (ICOL) processor leads to a decrease in APD. The C2R gave mostly low correlation coefficient values with a positive bias; APD ranged from 9 to 13% in the blue, whilst the poorest performing waveband was 620 nm with an APD of 67% compared to in situ data. The 3rd reprocessing with ICOL correction Ångström exponent showed the best correlation (R(sup2) = 0.951) with in situ data. The Chl-a product for case 1 waters from the 3rd reprocessing, Algal1, had the best agreement with in situ data, with a correlation coefficient of 0.796 and an APD of 54%. The Algal2 and C2R Chl-a products had low correlation coefficients and APD ranging from 72 to 103%. A dynamic per pixel classification technique for applying optimal MERIS Chl-a algorithms in the KwaZulu-Natal Bight or similar south-east African water types is described and evaluated in conjunction with the MERIS fluorescence line height product.