Bias in detrital zircon geochronology: a review of sampling and non-sampling errors

ABSTRACT

Detrital zircon (DZ) geochronology is based on a multistage, hierarchical sampling process in which sampling errors (statistical variance) and non-sampling errors (sampling bias) are endemic. Most of these errors are related to 1) initial collection of a non-random grab sample (typically from outcrop); 2) processing the grab sample to obtain a non-random grain-size sample of restricted grain size(s) (effectively a processed population); 3) randomly and/or non-randomly selecting DZs (often based on their size, shape, and colour) from the grain-size sample to obtain a hand-picked sample (effectively an analysed population); and 4) randomly and/or non-randomly radiometrically dating only certain DZs (due to limitations of instrument spot size and grain imperfections) from the hand-picked sample to obtain an analysed sample. Claims that hand-picked samples and analysed samples are randomly chosen are questionable, since these samples likely represent haphazard sampling attempting to emulate randomness. Non-randomness has major implications for calculations that determine sample size and inter-sample ‘sameness’ comparison, as both techniques generally depend on randomly chosen samples. The use of inter-sample comparison is particularly troublesome in that sameness is determined for the analysed sample, which is not representative of the initial grab sample. Yet more bias occurs with a ‘leap of inference’, the generalization of DZ ages from the analysed sample to the target population (the strata of interest) or to a general population (DZs derived from ultimate and proximate sources). Sampling bias can be reduced by unambiguously defining the sampled population (the population that inferential statistics apply), selecting representative samples (if possible, randomization should be used at every stage in the sampling process), and preserving the representativeness of subsamples by using correct sampling devices such as a microsplitter. It may be timely for standardized sampling, interpreting, and reporting guidelines so that detrital geochronology data are scientifically meaningful and replicable.

KEYWORDS:

• Detrital zircon
• geochronology
• sampling errors
• sampling bias

Acknowledgments

Dedicated to the memory of Don Murray ‘Trip’ Triplehorn (1934–2020), Professor Emeritus, University of Alaska Fairbanks, who shared his interests in the North, sedimentology, and dogs.

This manuscript benefited from thoughtful reviews by R. Leary, S. Puetz, and an anonymous reviewer, for which the author is grateful. The author also thanks R.J. Stern for editorial handling. Respectfully acknowledging that I live and work in the traditional territory of the Kwanlin Dünn First Nation and the Taa’an Kwächän Council of southern Yukon, Canada.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was funded by the author with no external support.