Mapping Surficial Sediment Distributions from Caicos Platform: A Quantitative Approach Integrating Statistical Analysis of Landsat Spectral Data and Field Observations

Stephen E. Kaczmarek¹ and Franciszek J. Hasiuk^2
1Upstream Research Company, ExxonMobil, Houston, TX
²Department of Geological Sciences, Universtiy of Michigan, Ann Arbor, MI

This investigation uses Landsat 7 spectral data in conjunction with cluster analysis algorithms and field-collected sediment sample data to create a surficial sediment texture map for the Caicos Platform. Cluster analysis utilizes statistical algorithms to group Landsat picture elements with similar spectral signatures into relatively homogeneous thematic classes. Carbonate sediment textures are subsequently assigned to each thematic class as part of a manual geologic interpretation that is calibrated using sediment data.

Our results indicate that the Caicos Platform is heavily grain-dominated. Geospatial calculations show that 7% of surficial sediments on Caicos are grainstone, 59% are mud-lean packstone, 18% are packstone, 5% are wackestone, and 1% are reef, with the remaining 10% characterized by exposed Pleistocene islands. Sediment distributions are highly asymmetric with most platform-scale facies variations following major depositional trends. At the 28.5 meter scale, the Landsat-derived facies map for Caicos demonstrates greater than 84% correlation with sediment data. This agreement indicates that the Landsat-derived facies map accurately characterizes the spatial dimensions and distribution of platform-scale depositional features, like grainstone shoals and tidal flats. Local-scale textural heterogeneity within individual depositional regimes is also identified.

Our observations suggest that surficial sediment distributions on the Caicos Platform are controlled by platform physiography and island orientation relative to the dominant hydrodynamic forces, such as the Antilles Current, easterly Trade Winds, and Atlantic swells. We also demonstrate that the coupling of statistical algorithms, Landsat data, and sediment data offers a powerful quantitative approach for investigating the spatial distribution of surficial sediments on modern carbonate platforms.

AAPG Search and Discovery Article #90078©2008 AAPG Annual Convention, San Antonio, Texas