Deciphering a Multipeak Event in a Noncomplex Set of Detrital Zircon U–Pb Ages
The determination of U–Pb ages from detrital zircons of sedimentary rocks using LA–ICP–MS has been widely used to develop studies of provenance analysis. A problem that frequently arises is to find a population that appears to be noncomplex despite several perceptible age peaks in their spectrum. These peaks are qualitatively defined through diagrams of relative probability (probability density function or PDF), but it is difficult to quantify their statistical significance relative to a zircon-forming multipeak event. Therefore, the question arises as to whether we can decipher and characterize a multipeak event in a noncomplex set of detrital zircon U–Pb ages. This work is an attempt to answer the above question by means of a statistical analysis. The objectives are: (1) to determine the most appropriate minimum number of zircon age populations (peaks); (2) to characterize each peak in terms of age and event duration; and (3) to compare results obtained for two datasets showing similar zircon ages. The process starts by using a cluster analysis to group zircon ages into a set of consistent clusters. A Gaussian kernel function is then fitted to each cluster and summed to obtain a theoretical PDF. At the end of the process, the best modelled PDF must coincide with the original PDF at ≥95 %, and the deciphered peaks can then be characterized.
KeywordsDetrital zircon data Population Peaks Comparisons of datasets Gaussian kernel function
This paper is a contribution to research project GONDWANA-PTDC/CTE-GIX/110426/2009 funded by Fundação para a Ciência e Tecnologia (Portugal). It is a contribution to project PEst-OE/CTE/UI4073/2014.