Retrieval chlorophyll-a concentration in the Taranto coastal area using remote sensed data

Monitoring of coastal waters for an early detection of pollution and eutrophication phenomena related to urban settlements and industrial activities is becoming of increasing importance for environmental protection and disaster prevention. For this purpose, high resolution, multi-spectral satellite data have been successfully applied in the detection of chemical, biological and physical processes developing in coastal waters, but they appear to be heavily penalized in the study of fast dynamical phenomena, due to their typically long revisiting time. At the same time, datasets from satellites for atmospheric and oceanic applications do have an appropriate revisiting time and an excellent spectral resolution, but generally are not considered suitable for coastal water studies, given their limited spatial resolution. In this paper, we present a preliminary evaluation of the utilization of oceanic sensors in the monitoring of coastal shallow waters. Among other parameters, we have measured the Chlorophyll-a concentration in the Taranto coastal waters, utilizing SeaWifS (SeaWide Field-of-view Sensor) data, as part of the SPICAMAR experiment conducted by environmental specialists, who have provided a very complete dataset relative to the ground (sea) truth. The "Taranto Seas" are three small water basins, nearly closed, with a total extension of about 55 km² and a small depth (7 m., 10 m. and 20 m. average depth respectively), located in southern Italy on the Ionian sea, very close to a large town and an important industrial settlement. Due to the low extension of the area of interest an important source of errors derived from pixels marginally contaminated by coastal land reflectivity. This effect has been partially eliminated trough a very accurate geo-referencing and selection of each single pixel to be used for the evaluation. A radiative transfer code has been used to remove the atmospheric effect by SeaWiFS image. To retrieval chlorophyll concentration we have implemented a semianalytic algorithms. The performance of these procedure and algorithms are presented to show the potential use and the limits of oceanic sensors for coastal waters monitoring.