ACROPORA CERVICORNIS ABUNDANCE ASSESSMENT AND HABITAT CHARACTERIZATION AT CORAL GARDENS, BELIZE USING IKONOS MULTISPECTRAL SATELLITE IMAGERY CLASSIFICATION

In recent decades Caribbean coral reefs have experienced drastic decline in live coral cover including one of the main framework-building coral species, Acropora cervicornis. Coral Gardens, Belize is one of the few remaining refuges with an abundant, healthy population of A. cervicornis. IKONOS multispectral satellite imagery of a 25 km² area near Ambergris Caye was analyzed to identify the live A. cervicornis cover in the greater Coral Gardens region. We used a supervised classification to separate A. cervicornis from other benthic cover such as mixed sand, turtle grass, and macroalgae. To capture the spectral character of A. cervicornis cover, a previously mapped refuge of live A. cervicornis located in the northern part of Coral Gardens was used for the classification “training areas.” 10 point features were then placed within the largest identified regions of suspected live A. cervicornis on the classified image. In the field, classification accuracy was tested by sending snorkelers to the region of suspected live A. cervicornis mapped around each point to document live coral cover, depth, orientation of live coral, species of corals present, and height of the tallest live coral. Points were recorded using a Trimble GPS unit for smaller areas of live A. cervicornis with an accompanying estimate of area coverage, and perimeters were mapped for larger stands of live A. cervicornis. Additionally, control points were also recorded to help refine the method of spectrally distinguishing live A. cervicornis from other benthic cover. Of the 10 predicted points, 7 had notable populations of A. cervicornis, leading to the conclusion that the initial classified image correctly identified the live A. cervicornis within the extent of the image, with a few false positive ID’s of turtle grass and coral rubble. Control points and newly mapped regions from the field data will be used in conjunction with different image processing methodologies to create a new classified map of the area. The new classification accuracy will be assessed noting false positives and false negatives in identifying A. cervicornis, and other benthic units will be mapped so that the area’s seafloor will be completely characterized. Identifying regions of A. cervicornis near Coral Gardens will be useful in continued monitoring of species health and abundance.