A Deterministic Approach to Geochemical Stratigraphy

Abstract

The scale and homogeneity of Large Igneous Provinces make them challenging to differentiate. The precision of modern geochemical analysis allows for the detection of heterogeneities within LIPS, but the act of deriving, classifying, and correlating geochemical stratigraphies remains qualitative and subject. In this study, I present three algorithms for the analysis of geochemical stratigraphies. stratify is an algorithm for dividing a geochemical section into units of comparable chemical similarity; stratClassify is a means for placing a series of samples into an existing, known stratigraphy; and stratMerge provides a means for comparing units between different stratigraphies and assessing the probability that they are equivalent. The potential of these algorithms is measured by their application to test cases in the Grande Ronde formation of the Columbia River Basalt and the Rajahmundry Traps. The algorithms stratify and stratClassify not only expose potential issues with the current stratigraphic definition of the Grande Ronde, but go so far as to provide a means for systematically deriving a stratigraphy in the area. stratMerge provides strong evidence that the Rajahmundry Traps are correlated with the Ambernali formation of the Deccan Traps, over 400 km away, a result which has consequences for volume estimates of the Deccan and methods of magma transport. These algorithms, having been tested briefly, are provided openly, with the intent they will be used and improved upon as geology embraces statistical rigour.