Validation of MERIS sensor's CoastColour algorithm for waters off the west coast of India

Chlorophyll-a (chl-a) retrieved using the MERIS CoastColour (CC) algorithm was evaluated for the coastal waters of the west coast of India, against in situ observations made as part of the Satellite Coastal and Oceanographic Research (SATCORE) programme. These observations include profiles of surface solar irradiance (Es) along with those of upwelling radiance and downwelling irradiance, measured using hyperspectral radiometry. Chl-a was also estimated from water samples. Furthermore, remote-sensing reflectance (R-rs) and chl-a were retrieved from MODIS-Aqua using the OC3M algorithm, and from MERIS using the OC4E algorithm. In addition, to understand the long-term seasonal variability, chl-a retrieved from the MERIS-CC algorithm was overlaid on monthly mean chl-a time series data from MODIS. Comparison of chl-a using MERIS-CC to that measured in situ showed wide scatter around the linear trend line. We observed that chl-a from MERIS-CC was underestimated for two-thirds of the observations, whereas with MODIS and MERIS it was 51% and 44%, respectively. Statistical analysis showed an improved performance in chl-a retrieval using the operational OC4E algorithm as compared to that of MERIS-CC. The time series analysis showed a good match between in situ chl-a and that derived from MODIS using the OC3M algorithm, whereas the MERIS-CC algorithm showed inconsistency in match-up with regard to both magnitude and trend. This inconsistency was more prominent during the low-chl-a scenario during the northern winter. We infer that algorithms such as OC4E and OC3M that use bands from the blue and green regions of the spectrum offer better chlorophyll retrieval in high-TSM or -CDOM concentration waters in comparison with CoastColour, which uses all bands across the spectrum.