Multivariate analysis of sedimentary environments using grain-size frequency distributions

Abstract

Multivariate statistics were used to characterize and test the effectiveness of grain-size frequencies as environmental discriminators. Sediment from the following two depositional systems along eastern Lake Michigan were studied: (1) a closed system with respect to available grain sizes (Little Sable Point), and (2) an open system (Sleeping Bear Point-Manitou Passage). Principal components analysis shows that grain-size distributions are composed of two or more subgroups that reflect surface creep bedload, mixed suspension bedload, and uniform suspension. Discriminant function and principal latent vector analyses of the Little Sable Point environments show that, when available sediment is limited with respect to grain size (0.5 φto 3.0 φ), similar size distributions can occur in environments supposedly characterized by different energy conditions. Sediment in the Sleeping Bear Point-Manitou Passage system is not restricted to available grain sizes and the environments discriminated very well (α < 0.001). The grain-size distributions are such that they reflect differences in energy conditions within the environments. It is apparent that the grain sizes available to a depositional system control to a great extent the effectiveness of environmental discrimination.