A harmonized radiocarbon chronology framework for subarctic North Pacific marine sediment cores

Abstract. Published marine sediment chronologies from the subarctic North Pacific differ in calibration curves, reservoir corrections, input selection, and modelling conventions, complicating comparisons among palaeoceanographic records. We present a chronology framework for 17 cores from the Bering Sea, Gulf of Alaska, Sea of Okhotsk, and open subarctic North Pacific. The main inventory contains 19 core or segment records: 16 Marine20–rbacon reconstructions, two separately identified legacy calendar-age models, and one segment without a chronology. The 18 canonical age-model tables provide posterior median ages and pointwise 95 % intervals on a standardized 201-depth grid. Users can assign these ages to depth-registered proxy measurements from the same cores after verifying depth-scale correspondence; the framework cannot date a new core without its own chronological constraints.

The archive preserves 555 chronology-control rows with their provenance, material roles, and admission decisions. The 18 verified formal model-input files contain 285 rows representing 282 dated-depth clusters or legacy calendar-age nodes. Twenty-eight constraints misassigned to surface chronology during an earlier stage of this compilation are retained with their material-role and exclusion records, allowing users to inspect or revise the admission decisions. The primary precision descriptor is the median width of the pointwise 95 % posterior age interval within each dated interval. Among 14 inventory rows assigned a derived convenience bin, this statistic ranges from 677 to 2765 yr, with a median of approximately 1100 yr. Sampling diagnostics are reported at all evaluation depths. Independent reruns of four preselected records changed this continuous statistic by at most 6.64 yr, although one boundary-adjacent record crossed a derived-bin boundary.

Posterior precision is conditional on calibration, the reservoir correction and its adopted date-level uncertainty, input representation, and the modelling protocol; it does not establish absolute chronological accuracy. The adopted reservoir corrections are based on modern observations, whereas deglacial reservoir histories remain incompletely constrained. The release combines age tables, audited inputs, cluster definitions, diagnostics, per-record JSON/CSV packages, source references, and implementation records to support reuse and alternative chronological modelling. Executed model inputs are preserved, but recovery of the full source-to-prepared transformation chain is incomplete. The data are deposited at PANGAEA under https://doi.pangaea.de/10.1594/PANGAEA.996868.