Multivariate analysis of ichnofossil associations in the subsurface Bluesky Formation (Albian, Alberta, Canada)

Abstract

The quantitative analysis of ichnological associations is in an early stage of development. Several studies of spatial distributions of ichnofossils have provided insight into spatial ecological interactions in modern and ancient environments. This study examines another dimension, the development of paleoecological and paleoenvironmental conditions through time. Ichnological data from diamond drill cores of 22 wells that penetrated the Bluesky Formation of northwestern Alberta were digitised. Q-mode cluster analysis distinguished 8 ichnocoenoses based on associations of ichnofossils and a few diagnostic physical structures. R-mode cluster analysis of these ichnofossil associations distinguished 3 major groups. The control provided by an existing subjective classification of the original facies intervals into trace fossil suites allowed an examination of the efficacy of several different clustering functions. Ward's method utilising Euclidian distance for the similarity coefficients and the error sum of squares for the fusion transformation function appeared to be the optimum technique for this study. Two methods of Markov analysis employed the 8 ichnocoenoses as transition states. These were the Gingerich-Read row-scaling method, using Harper's binomial probabilities as a test of significance, and the Goodman iterative proportional fitting method to expected transition frequencies, using the χ 2 statistic and Türk's standardised residuals as a test of significance. The iterative proportional fitting technique showed significant deviation from independence at the 95% confidence limit. Both methods distinguished similar significant biofacies transitions. The resultant ichnocoenosis model for the Bluesky Formation in the study area is that of a nearshore to marginal regressive sequence following a sudden transgression over a coastal plain or swamp environment.