A multispectral classification scheme was developed to detect the cyanobacteria Trichodesmium spp. in satellite data of the sea-viewing wide field-of-view sensor (SeaWiFS). The criteria for this scheme were established from spectral characteristics derived from (1) SeaWiFS imagery of a Trichodesmium bloom located in the South Atlantic Bight and (2) modeled remote sensing reflectances of Trichodesmium and other phytoplankton. The classification scheme, which is valid for moderate chlorophyll concentrations of Trichodesmium in coastal waters, is based on the magnitude of the 490-channel reflectance and the spectral shape of remote sensing reflectance at 443, 490 and 555 urn. Analysis suggests that the spatial structure of Trichodesmium populations at sub-pixel scales must be considered when employing spectral characteristics to detect their presence in satellite imagery. This study demonstrates the potential of mapping Trichodesmium from space using spectral observations, even in waters as optically complex as the South Atlantic Bight. Future efforts, which will incorporate ancillary data such as wind speeds and water temperature, will improve the likelihood of correct identification. (C) 2001 Elsevier Science Ltd. All rights reserved.
507CETimes Cited:38Cited References Count:26