Grain selection for representative detrital zircon age populations

Detrital mineral analysis is deployed to address a diversity of Earth science questions, e.g., sediment provenance through zircon U–Pb age distributions. However, conventional approaches in sample preparation and grain selection involve inevitable sources of geological and laboratory bias that can distort the relative abundances of distinct age modes and thus skew geological interpretations. It has been established that detrital zircon grain shape reflects the combined result of conditions during primary formation and subsequent modification through Earth surface processes. Here, we explore if the population structure of 2D grain dimensions can be leveraged as a sub-sampling tool to produce more representative selections of analyses at lower sample numbers. We use two comparable detrital zircon populations with > 1500 concordant zircon U–Pb analyses each from distinct depositional environments (fluvial vs. shoreline) with differing degrees of sorting and complexity in source contributions (as implied by U–Pb age structures). Our results demonstrate that selection of the largest – akin to human operator bias – or smallest zircon grains can result in statistically distinct populations until > 1000 grains are chosen. Preserving a representative grain size distribution by sub-sampling to the grain shape population's mean value can circumvent this deleterious effect, irrespective of the degree of sorting. Results of ca. 600,000,000 model iterations imply that age population sampling variations fall within typical facies variability at a single stratigraphic level when analyzing ca. 240 grains and within the typical variability of geological formations when analyzing only ca. 70 grains using a representative selection strategy.