Dataset DOI: 10.5061/dryad.xgxd254wx

Description of the data and file structure

This dataset contains four CSV files (Supplementary Data 1–4) and one MATLAB code file (Supplementary Code 1) supporting the findings of the paper: Dong, X. et al., Glacial abrupt transitions in ocean circulation were much slower than ice-core records imply.

Note: The original Excel files (Supplementary_Data_1.xlsx, Supplementary_Data_2.xlsx, Supplementary_Data_3.xlsx and Supplementary_Data_4.xlsx) are also provided alongside the CSV files for readers who wish to consult the original formatted data.

The dataset includes U-Th dating results for five speleothems, d18O time-series and annual lamina thickness data, Greenland ice-core records anchored to U-Th chronologies, stacked speleothem and ice-core records, sensitivity test results, and MATLAB code for annual lamina counting.

Files and variables

File: Supplementary_Code_1.m

Description: MATLAB code for annual lamina counting

File: Supplementary_Data_1.csv

Description: U-Th dates for 5 speleothems (Cherrapunji-2, Cherrapunji-2017-1, SX-1, DSB-3, MAW13-6)

General notes on U-Th dating:

• Decay constants used:
  • U decay: λ₂₃₈ = 1.55125×10 ^-10 (Jaffey et al., 1971)
  • U decay: λ₂₃₄ = 2.82206×10 ^-6 (Cheng et al., 2013 EPSL)
  • Th decay: λ₂₃₀ = 9.1705×10 ^-6 (Cheng et al., 2013 EPSL)
• Definitions:
  • d234U* = ([^234^U/^238^U]activity – 1) × 1000
  • d234UInitial was calculated based on the ²³⁰Th age (T) as: d234UInitial = d234Umeasured × e^(λ₂₃₄ × T)
• Correction for initial ²³⁰Th:
  • Corrected ²³⁰Th ages assume an initial ²³⁰Th/²³²Th atomic ratio of (4.4 ± 2.2) × 10 ^-6, which corresponds to a material at secular equilibrium with a bulk Earth ²³²Th/²³⁸U value of 3.8. The errors are arbitrarily assumed to be 50 %.
• Age convention:
  • B.P. stands for "Before Present", where the "Present" is defined as the year 1950 C.E.
• Notes on specific samples:
  • SX-1: The sample was counted separately in three sections: 520–554 mm, 574–620 mm, and 640–670 mm. Therefore, the band numbers in each section correspond only to the counted bands within that specific section.
  • Cherrapunji-2: Samples Cherrapunji-2-829 and Cherrapunji-2-849 were not used in the final age model construction (see methods).

File: Supplementary_Data_2.csv

Description: d18O time-series for all 5 speleothems, plus annual lamina thickness for Cherrapunji-2 and SX-1

Notes on age models:

• For Cherrapunji-2, the age model for the whole section is based on a combination of annual lamina counting and ²³⁰Th dates (see main text).
• For SX-1, the age model for certain periods is also based on the combination of annual lamina counting and ²³⁰Th dates. In addition, we provide a full StalAge-based age model for the entire SX-1 record (see methods).
• For all other speleothem records (Cherrapunji-2017-1, MAW13-6, DSB-3), the age models were constructed using the StalAge algorithm.

File: Supplementary_Data_3.csv

Description: Greenland ice-core records (NGRIP, GISP2, GRIP, NEEM, DYE3) on U-Th-based chronologies

General notes:

• New age (years BP): This column represents the new Greenland ice-core chronology ages, anchored using the U-Th dated speleothem records. The original GICC05 chronology ages (in years b2k or BP) are also provided for reference.
• All ice-core d18O data are reported in ‰ (per mil) versus VSMOW.

File: Supplementary_Data_4.csv

Description: Stacked records and sensitivity tests

General notes:

• This file contains the data used to generate Figures 2h, 2i, and 2j in the main text, as well as the results of sensitivity tests.
• It also includes stacks of NGRIP ice-core d18O and Antarctic ice-core dln records for Dansgaard-Oeschger (DO) events 1 through 13 for comparison.

Code/software

To run the lamina counting code:

1. Open Supplementary_Code_1.m in MATLAB.
2. Edit Line 29 to specify the input '.csv' filename (exported from the ImageJ software).
3. Edit Line 34 to adjust peak detection parameters ('MinPeakDistance' and 'MinPeakProminence').
4. The code will calculate the number of laminae and the annual lamina thickness.

See paper Supplementary Text 1.4 for method details.