Abstract
In this study bio-optical water quality indicators, chlorophyll a , colored dissolved organic matter (CDOM), and total suspended matter (TSM) were derived from the Envisat-MERIS satellite data and were compared with in situ measurements collected in the Lithuanian optically Case 2 coastal waters of the Baltic Sea. Eight MERIS full-resolution Level 1b images, acquired during late spring and summer 2010, were processed using five, neural network-based processors for optically Case 2 or coastal and inland waters: FUB, C2R, Eutrophic, Boreal, and standard MERIS Level 2. Results showed that the FUB processor provided the most accurate estimates of the concentration of chlorophyll a [R 2 =69% ; mean absolute error(MAE)=7.76 mg/m 3 ] and TSM (R 2 =89% ; MAE=3.93 g/m 3 ). In situ CDOM absorption was most accurately estimated using the Boreal processor (R 2 =69% ; MAE=0.20 1/m ). We analyzed the factors that were most influential in explaining the differences in the accuracy and found that the Secchi depth and the sampling time were the most important factors. The greatest differences between satellite-derived and in situ values of water quality indicators were in correspondence with the lowest Secchi depth, suggesting that the plume zone created by freshwater coming from the hyper-eutrophic lagoon was the most sensitive region for the validation. The evident match between in situ measurements and satellite-based estimates was observed when field measurements were acquired 1-2 h before to approximately 2-4 h after the satellite overpass. Results of this validation work confirmed that remote sensing techniques are suitable for monitoring the changes of optical constituents in Lithuanian coastal waters.
