Abstract
Discriminant function analysis is an effective tool for differentiation of strand-line environments, represented by calcareous sediment from the Bermuda Islands. Sediment samples from the beaches of Bermuda were separated on the basis of wave energy regime and tidal position: subtidal, intertidal, and supratidal (dune and eolian flat). Sediment from the beaches that face the lagoon is composed of fragments of (1) the green alga, Halimeda, (2) bivalves, and (3) Pleistocene, skeletal eolianite. Sediment from beaches that face the open ocean is composed of fragments of (1) the Foraminifera, Homotrema, (2) coral, (3) coralline algae, and (4) Pleistocene, skeletal eolianite. Differences in sediment texture and ability to discriminate depositional environments are related to skeletal composition of the sediment and to wave-energy regime. The combination of phi mean, phi standard deviation, skewness, and kurtosis is most efficient for separating all categories of strand sediment by discriminant analysis. Phi mean and standard deviation combined are also efficient variables for discrimination. Skewness and kurtosis are significantly lower in overall discriminant efficiency, and neither can be utilized successfully when paired with only one other textural variable. Different discriminant functions must be used on the samples from each shore to obtain optimum discrimination, which suggests that specific textural discrimination criteria may be locally, but not generally, valid.
