Monday 22 August 2005

PB1

PM0082
Hyperspectral sensing of coastal waters
Vargas, Marco¹, Gross, Barry¹, Moshary, Fred¹, Ahmed, Samir¹
1 The City College Of The City University Of New York, USA
The optical properties of natural bodies of water are influenced by three main components. Phytoplankton and other microscopic organisms, suspended inorganic material, and dissolved organic substances from decaying organic matter. Whereas the optical properties of deep off shore waters (case I waters) are primarily a function of phytoplankton concentration, coastal and inland waters represent a more complex optical environment. The optical properties in these waters (case II waters) are due to a mix of phytoplankton, total suspended sediments (TSS), and chromophoric dissolved organic matter (CDOM) which can vary independently from each other. In case I waters, where the main biooptical activity is phytoplankton concentration and in particular, chlorophyll pigments, algorithms such as those on SeaWIFS ocean color satellite sensor utilizing the blue-green wavelength band ratios, which probe the main absorption band of chlorophyll, are sufficient. Unfortunately, these algorithms cannot be used in turbid waters due to the interference in the optical signal from TSS and CDOM. In this paper, we examine an alternative approach which utilizes a set of close wavelengths within the second absorption band of chlorophyll in the near infrared (NIR) between 670-700 nm. For closely spaced wavelengths, TSS will not result in a significant wavelength dependence of the backscatter spectrum. Also, in this range, the impact of CDOM (primarily in terms of absorption) is rather small. We show the robustness of this approach in coastal waters using simulations based on biooptical modeling, as well as insitu measurements in Long Island Sound. On the other hand, when bottom affects are introduced for shallow waters, hyperspectral measurements are required to obtain simultaneously ocean water parameters and ocean bottom parameters. We show the potential retrieval improvement from hyperspectral sensors such as the planned GOES-R HES instrument over multispectral sensors such as SeaWIFS and MODIS.

Return to Poster Presentations