An Evaluation of the Accuracy of Meris Radiometric and Geophysical Products for a Northern Adriatic Sea Coastal Site

The MERIS radiometric and geophysical products are assessed using in situ data collected at the Acqua Alta Oceanographic Tower in the northern Adriatic Sea from May 2002 to May 2003. The analysis is restricted to satellite and in situ data taken within ±1 hour difference to minimize the effects of the spatial and temporal non-homogeneity of atmosphere and sea around the measurement site. The comparison analysis between satellite and in situ normalized water leaving radiances for 28 match ups exhibits spectrally averaged mean differences of 6% in keeping with match-ups analysis of SeaWiFS and MODIS data for the same site. When the analysis is restricted to MERIS Case2_S flagged data (i.e., turbid water data), the comparison over 16 match-ups shows spectrally averaged mean differences of 19.7%. Differently when the analysis is restricted to MERIS data non-flagged as Case2_S, the comparison over 12 match-ups shows spectrally averaged mean differences of -14.9%. These systematic biases of satellite data versus in situ measurements suggest different accuracies in the atmospheric correction of data taken over turbid and non-turbid waters. In addition, occasional negative water leaving radiances at 665 nm (approximately 1/3 of the cases) for non-turbid water data, suggests an overcorrection of the atmospheric effects. The comparison between MERIS derived pigment concentration (i.e., Algal I product) and total chlorophyll a determined from in situ samples analyzed through High Performance Liquid Chromatography or estimated by applying a regional algorithm to in situ radiometric data, shows an overestimate of a factor 2 for the satellite products regardless of the water turbidity. This result could be mostly justified by the application of an ocean color algorithm not suitable for the northern Adriatic Sea coastal waters. In fact when applying a regional band ratio algorithm to MERIS radiometric data the agreement with in situ measurement significantly improves. The comparison of MERIS derived TSM concentration with in situ values determined from gravimetric analysis of samples or obtained by applying a regional algorithm to in situ radiometric data, do not show any correlation. When applying a regional band ratio algorithm to MERIS radiometric data the agreement significantly improves for the turbid water data only. This result is justified by the too low water leaving radiance data obtained in the red (due to an overcorrection of the atmospheric effects) and applied for estimating TSM with the chosen algorithm.