Estimating best fit binary mixing lines in the Day plot

Abstract

The Day plot is employed commonly to make inferences concerning the mean domain state of magnetic mineral assemblages. In recent years the concept of the Day plot as a representation of a two-part (binary) mixing space has gained popularity, with data sets being compared to theoretical curves that represent combinations of particles with different domain states. Our understanding of the behavior of mixtures within the Day plot, however, remains limited, and little progress has been made in terms of quantitative statistical analysis. We present an approach, based on linear mixing theory, with which a data-optimized binary mixing line can be found for a collection of hysteresis loops and their corresponding backfield demagnetization curves. The empirical best fit mixing line can then be used to determine a trend through the Day plot. Such trend lines help to constrain interpretation of rock magnetic data and aid in identification of binary mixtures. In cases where the trend line does not provide a good fit to the data, it can be concluded that the examined samples do not originate from a two-part linear mixing. The proposed method is demonstrated using both numerical simulations and geological data sets.