An evaluation of remote sensing techniques for enhanced detection of the toxic dinoflagellate, Karenia brevis

Abstract

Optical techniques were investigated to enhance current bloom detection capabilities in support of an operational system for forecasting harmful Karenia brevis blooms along the west coast of Florida, within the Gulf of Mexico. Algorithms pertaining to backscatter and changes in spectral shape of remote-sensing reflectance were applied to SeaWiFS and MODIS imagery during known K. brevis and non-K. brevis events. A method to remove resuspended chlorophyll in Texas showed limited use when applied to several scenes following tropical storms off the west Florida coast. This analysis suggests that an ensemble image approach, wherein a combination of a chlorophyll anomaly, spectral shape at 490 nm and a backscatter ratio product would provide an improvement in satellite detection of K. brevis blooms. For southwest Florida, the combination of these methods through an ensemble approach may lead to an increase in user accuracy by 30–50%, as a result of correctly identifying non-K. brevis features. Where available, MODIS FLH scenes were analyzed to determine their use in K. brevis detection. However, insufficient imagery was available to make a fair assessment. Similar approaches could be applied to bloom tracking and monitoring in other regions.