Benthic sediment facies maps were constructed for a series of large, isolated carbonate platforms, including (1) Great and Little Bahamas Banks; (2) Caicos Platform, British West Indies; (3) Chinchorro Bank, Mexico; (4) Glovers Reef, Belize; (5) south Cocos (Keeling) atoll, Indian Ocean; and (6) Bu Tini shoal, Persian Gulf. Facies maps were generated by applying statistics-based image analysis algorithms, called unsupervised classifications, to Landsat 7 multispectral satellite data. Classification results were subsequently validated with sediment data to create geologic facies maps.

Landsat-derived facies maps demonstrate 82–85% agreement when compared to sediment data. At the platform scale, Landsat-derived facies maps accurately capture the principal depositional facies observed on each platform and are in general agreement with published maps generated through conventional mapping techniques. Examination at more detailed scales reveals that Landsat-derived maps differ from conventional maps with respect to the spatial dimensions and shapes of facies bodies. Landsat-derived maps show a level of complexity and heterogeneity that is more realistic than shown in previously published maps, which are characterized by broad, homogeneous facies belts. These results suggest that application of statistical algorithms to Landsat data, coupled with sediment data, provides a cost- and time-efficient method for quantitatively mapping spatial variability of depositional facies in modern carbonate environments. Landsat-derived facies maps, like the ones presented here, provide depositional analogs for subsurface carbonate reservoirs as well as a global data set for extracting predictive relationships between the occurrence and distribution of carbonate sediments that can aid in global hydrocarbon exploration and production.