Tracing Geological Time in Western Australia: Coupled Ar-Ar and In Situ Rb-Sr Geochronology

Abstract

A persistent misconception in geochronology is that a rock has a single definitive age. Many isotopic systems instead record thermal histories shaped by multiple tectonothermal events, including partial re-equilibration, recrystallisation and episodic fluid interaction. Such histories challenge statistical models that assume a single true age with Gaussian analytical scatter.

Six samples from the Paterson Orogen (Yeneena Basin and Rudall Province), at the eastern margin of the Pilbara Craton, and one from the eastern margin of the Yilgarn Craton were analysed. Both areas are geologically underexplored. Furnace step-heating Ar-Ar geochronology of mineral separates, enabling quantitative temperature-time modelling, was combined with in situ Rb-Sr geochronology by LA-ICP-MS/MS.

Yeneena samples exhibit relatively homogeneous Ar degassing patterns across analysed minerals, yielding Ar-Ar ages of 640-590 Ma, consistent with the Paterson Orogeny and local granitic intrusions. Rudall Province samples show more variable patterns and dispersed ages of 880-590 Ma. In situ Rb-Sr data were processed using different approaches, including the open-source software Kasper Julia (Vermeesch, 2025b). A standardised workflow yields dates of 1380-590 Ma for Paterson Orogen samples. Rudall Province samples show overdispersion associated with differential retention of geochronological information by minerals such as biotite and white mica.

Neutral terms are proposed for investigating overdispersed datasets: "variochron" describes an overdispersed dataset, while "divergochron" describes a variochron with clear multi-generational signatures. These provide alternatives to "errorchron", criticised as pejorative (Vermeesch, 2025a). Geochemical decomposition of mineral populations helps interrogate divergochrons: combining isotopically distinct biotite and white mica can generate spurious "ghost" ages, whereas separate analysis recovers geologically meaningful age components.

A Bayesian approach incorporating prior geological information and accommodating uncertainty and multi-component, non-Gaussian age distributions is used to assess the timing and existence of the proposed "Miles Orogeny/Event" (840-654 Ma), which lacks robust chronological and tectonic evidence. An adaptive Student-t mixture model allows model complexity to respond to the data while remaining robust to outliers. Model-Averaged posterior predictive checks indicate that two components adequately describe the data: the Paterson Orogeny and a Mesoproterozoic event likely linked to the Parnngurr Orogeny. A distinct third "Miles Event" lacks statistical support in these data, indicating that it is merely an artefact of overfitting poorly constrained, sparse geochronological data rather than a "real" tectono-thermal event. Together with microstructural information preserved by in situ analysis, Ar-Ar constraints and the regional geological context, these results favour interpreting purported "Miles" ages as a composite apparent-age pattern arising from partial isotopic resetting, unresolved mineral mixing and the conflation of ages recording unrelated processes, rather than as evidence for a coherent regional geological event. This interpretation accounts for the pronounced overdispersion of "Miles" ages observed across multiple isotopic systems and datasets. Show more