Removing Complications of a Dominant Broad Age Spectrum Source: A New Approach to the Mixing of Detrital Zircons.

Detrital zircons are often considered the panacea of provenance analysis but are often limited to inquiry based solely on the presence or absence of populations and thus restricted to qualitative comparisons. However, if one broad age spectrum source dominates a detrital dataset other sources may be obscured. We present a method for removing a broad age spectrum source through a process of randomly drawing grains from the dominant source, pairing them with sample grains and removing matched aged grains. This results in a residual population representing all remaining unassociated ages, which can then be compared to other potential sources using PDP crossplot and likeness comparisons. The sources best matching our post-removal sample are resampled in varying proportions, with the dominant, broad age spectrum source, to recreate the pre-subtraction sample by maximizing the population comparison tests.

We test our method with a synthetic dataset of 18, 10,000-grain 'sandstone' populations created from 6 easily differentiable sources and 1 complex source representing the dominant, broad age spectrum source. Synthetic detrital zircon age populations are created by randomly selecting from each of the synthetic populations. This method was able to consistently determine the correct sources and their mixing proportions to within ±10% of the original artificial sandstone proportions. Covariance between the two comparison tests also shows improvement relative to their pre-removal counterparts. Previously published detrital zircon age data from modern sediment samples of the Yangtze River (He et al., 2013) show a dominance of ages commonly attributed to the Triassic Songpan Ganze Terrane, which has a broad age spectrum. We applied our method to these Yangtze age populations in conjunction with 35 potential sources units compiled and consolidated from the literature. Applying this technique to modern Chinese sediments identifies previously unknown contributing sources and their mixing proportions, which can significantly increase our ability to examine spatial and temporal changes and patterns often sought in provenance studies.