Sedimentary successions preserve the imprint of climatic, geochemical, and ecological change across Earth history, yet extracting high–precision temporal information from these archives remains a central challenge for integrated stratigraphy. Classical astrochronologic approaches typically resolve time only to the scale of precession, limiting our ability to interrogate the rates and durations of environmental perturbations that shaped the geological past.
High–resolution X–ray fluorescence (XRF) core scanning potentially offers an opportunity to overcome these constraints by providing 20–30 elemental series at ~0.2 mm spacing across intervals spanning numerous Milankovitch cycles. These multidimensional datasets capture coherent astronomical pacing signals embedded within chemically diverse sedimentary components, opening the door to a new generation of stratigraphic tools.
Here, we present a multidimensional cyclostratigraphic framework designed to advance the stratigraphic and paleoenvironmental toolbox. Our new cyclostratographic algorithm, AstroComb, employs a probabilistic, covariance–based approach to detect Milankovitch periodicities across multiple elemental series and to quantify uncertainty in inferred sedimentation accumulation rates. Building on such lower–resolution astrochronologic models, we present a second algorithm, ProBE4T (pronounced "Pro Beat"), which integrates these astronomical constraints with lithotype–specific sedimentation behaviour, inferred through unsupervised clustering of elemental and mineralogical estimates and Bayesian inversion under total–duration constraints. This probabilistic workflow distributes time across sedimentary successions at the resolution of the geochemical data itself and explicitly tests the hypothesis that chemically distinct lithotypes accumulate at distinct rates, thereby extending age–model refinement beyond the conventional precession limit.
The resulting age–depth models reveal substantial heterogeneity in time recorded per unit thickness, enabling precise temporal localization of paleoenvironmental signals, such as rapid climatic events, shifts in geochemical cycling, and changes in oceanic redox structure applicable from the Archean to the Holocene. By leveraging the full multidimensionality of XRF data and embedding probabilistic inference at each step, this approach expands the range of sedimentary archives amenable to high–resolution sediment accumulation rate determination and provides a generalizable methodology for integrating geochemical, lithological, and astronomical information.
The resulting age–depth models reveal substantial heterogeneity in time recorded per unit thickness, enabling precise temporal localization of paleoenvironmental signals, including rapid climatic events, shifts in geochemical cycling, and changes in oceanic redox structure, from the Archean to the Holocene. By leveraging the full multidimensionality of XRF data and embedding probabilistic inference at each step, this approach expands the range of sedimentary archives amenable to high–resolution sediment accumulation rate determination and provides a generalizable methodology for integrating geochemical, lithological, and astronomical information. The ProBE4T framework enhances our ability to explore temporal variability with uncertainties in sedimentary archives, opening new avenues for investigating climatic, geochemical, and ecological change at finer temporal scales across Earth history.