Water quality monitoring by thematic mapper in coastal environments. A performance analysis of local biooptical algorithms and atmospheric correction procedures

Abstract

The Thematic Mapper (TM) sensor's ability to monitor localized phenomena associated with anthropogenic pollution in coastal waters has been investigated with reference to experimental data collected on a test site in the Gulf of Naples. In situ measurements yielded the biooptical algorithms for chlorophyll and sediment retrieval as well as an optical model for the computation of the subsurface reflectance as a function of water composition, specific to the site. The algorithms' performance was assessed through a sensitivity analysis, taking into account the correlation among chlorophyll, sediment, and yellow substance contents, as well as the stratification characteristics of the water body. Sensitivity analysis techniques were similarly used to evaluate the performance in the coastal environment of three atmospheric correction models, which assume uniform aerosol load and composition, uniform aerosol composition, and varying aerosol load and composition, respectively. A comparison with sea-truth data collected for the Landsat overpass of 22 June 1988 showed a satisfactory correspondence between measured concentrations and concentrations retrieved from the TM data using the above atmospheric correction procedures and “local” biooptical algorithms. Consistent with the indications of the sensitivity analysis, the best agreement was obtained by the atmospheric model assuming varying aerosol load and composition.