Skip all navigation and jump to content Jump to site navigation
About MODIS News Data Tools /images2 Science Team Science Team Science Team

   + Home
MODIS Publications Link
MODIS Presentations Link
MODIS Biographies Link
MODIS Science Team Meetings Link



McKinna, LIW; Furnas, MJ; Ridd, PV (2011). A simple, binary classification algorithm for the detection of Trichodesmium spp. within the Great Barrier Reef using MODIS imagery. LIMNOLOGY AND OCEANOGRAPHY-METHODS, 9, 50-66.

A binary classification algorithm to detect the presence of high surface concentrations of the nitrogen-fixing cyanobacterium Trichodesmium spp. was developed for high spatial resolution (250 m) imagery of the Moderate-resolution Imaging Spectroradiometer (MODIS). Above-water hyperspectral radiometric measurements of dense Trichodesmium surface aggregations (>10 mu g/L Chlorophyll a) showed that the water-leaving radiance L(w) at wavelengths greater than 700 nm were much higher in magnitude (>0.05 W m(-2)sr(-1)) relative to the visible wavelengths 400-700 nm (<0.03 W m(-2)sr(-1)). The binary classification algorithm is based on three criteria. The first criteria relied on the difference in magnitude between the MODIS normalized water-leaving radiance (nLw) at the 859 and 678 nm wavebands. The magnitude of the nLw at the 555 and 645 nm wavebands relative to nLw 678 nm waveband formed the second and third criteria, respectively. The classification algorithm was tested on a small subset of 13 MODIS images with corresponding Trichodesmium sea-truths and yielded an 85% accuracy. Fine-scale features consistent with dense Trichodesmium surface aggregations, such as eddy swirls and windrows, appear to be well represented with the algorithm results. The algorithm was also found to be robust in the presence of highly reflective, potential confounding effects.



NASA Home Page Goddard Space Flight Center Home Page