A multivariate analytical strategy for classifying paleoenvironments

A multivariate analytical strategy is proposed for aiding the investigator in extracting maximum information from environmental data. Data are carefully coded and scaled and are tested for redundancy using R-mode cluster analysis. The samples are partitioned into environmental classes using Q-mode cluster analysis. Q-mode ordination facilitates interpretations, which usually can be verified by comparison with field relationships. Discriminant analysis serves as an identification procedure for extending the classification to unknown samples. The strategy is demonstrated by application to Cape Hatteras microorganism distributions and Devonian sedimentary facies.