DISCRIMINATION OF CORAL REEFS, SEAGRASS MEADOWS, AND SAND BOTTOM TYPES FROM SPACE - A DOMINICAN-REPUBLIC CASE-STUDY

In order to monitor changes caused by local and global human actions to a coral reef ecosystem, we ''sea-truthed'' a natural color Landsat Thematic Mapper (TM) image prepared for a coastal region of the northwestern Dominican Republic and recorded average water depth, precise geographical positions, and bottom types (seagrass, 15 sites; coral reef, ten sites; and sand, six sites). There were no significant differences in depth for the bottom type groups. The depths ranged from 0 to 16.1 m (0 to 52.7 ft). Sand > seagrass > coral in mean Landsat digital counts (proportionol to radiance) for the three Landsat TM visible bands (TM 1, TM 2, TM 3); sand bottom sites had significantly greater digital counts than seagrass and coral sites in TM 1 only. Mean digital counts of seagrass and coral reef sites did not differ significantly in any hand. A multivariate analysis of variance using all three bands gave similar results. A ratio of the green/blue bands (TM 2/TM 1) showed there was a spectral shift associated with increasing depth but not bottom type. Due to small-scale patchiness (< 30 m by 30 m), seagrass and coral areas were difficult to distinguish, but sandy areas can be distinguished using Landsat TM imagery and our methods.