Matrix profile analysis of Dansgaard-Oeschger events in palaeoclimate time series

This dataset includes all the datafiles and computational notebooks required to reproduce the work reported in the paper “Characterisation of Dansgaard-Oeschger events in palaeoclimate time series using the Matrix Profile”:

Input datafiles

  • time series (20-years resolution) of oxygen isotope ratios (δ18O) from NGRIP ice core on the GICC05 time scale (source: https://www.iceandclimate.nbi.ku.dk, DOI: 10.1016/j.quascirev.2014.09.007): the 1st columns is the time in ka (10³ years) b2k (before A.D. 2000), and the 2nd column the oxygen isotope concentration;

  • time series (20-years resolution) of calcium concentration (Ca2+) from NGRIP ice core on the GICC05 time scale (source: https://www.iceandclimate.nbi.ku.dk, DOI: 10.1016/j.quascirev.2014.09.007): the 1st columns is the time in ka (10³ years) b2k (before A.D. 2000), and the 2nd column the Ca2+ concentration;

  • time series (20-years resolution) of calcium concentration (Ca2+) from NGRIP ice core on the GICC05 times scale, artificially shifted by 10 ka (500 data points): the 1st columns is the time in ka (10³ years) b2k (before A.D. 2000), and the 2nd column the Ca2+ concentration;

  • time series (20-years resolution) of calcium concentration (Ca2+) from NGRIP ice core on the GICC05 times scale, trimmed by 10 ka (500 data points): the 1st columns is the time in ka (10³ years) b2k (before A.D. 2000), and the 2nd column the Ca2+ concentration;

Code and computational notebooks

  • R code for visualisation of matrix profile calculations;

  • jupyter notebook (python) containing the matrix profile analysis of the oxygen isotope time series;

  • jupyter notebook (python) containing the matrix profile analysis of the calcium time series;

  • jupyter notebook (python) containing the join matrix profile analysis of oxygen isotope and calcium time series;

  • jupyter notebook (R) for visualisation of matrix profile results of the oxygen isotope time series;

  • jupyter notebook (R) for visualisation of matrix profile results of the calcium time series;

  • jupyter notebook (R) for visualisation of join matrix profile results;

Output datafiles

  • matrix profile of the oxygen isotope time series (sub-sequence length of 2,500 years): the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence);

  • matrix profile of the oxygen isotope time series (sub-sequence length of 3,000 years): the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence);

  • matrix profile of the oxygen isotope time series (sub-sequence length of 3,500 years): the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence);

  • matrix profile of the calcium time series (sub-sequence length of 2,500 years): the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence);

  • join matrix profile for the oxygen isotope and calcium records: the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence); the 3rd and 4th columns are included in the default output but are only valid in the case of self-joins, and thus here not applicable (flagged with a constant negative value of -1);

  • join matrix profile for the oxygen isotope and shifted calcium records: the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence); the 3rd and 4th columns are included in the default output but are only valid in the case of self-joins, and thus here not applicable (flagged with a constant negative value of -1);

  • join matrix profile for the oxygen isotope and trimmed calcium records: the 1st column contains the matrix profile value (distance to the nearest sub-sequence), the 2nd column contains the profile index (the zero-based index location of the nearest sub-sequence); the 3rd and 4th columns are included in the default output but are only valid in the case of self-joins, and thus here not applicable (flagged with a constant negative value of -1);

Data og ressourcer

Yderligere info

Felt Værdi
Forfatter Susana Barbosa, Maria Eduarda Silva, Denis-Didier Rousseau
Last Updated maj 13, 2024, 10:08 (UTC)
Oprettet februar 6, 2024, 10:09 (UTC)
Citation Barbosa, S., Silva, M.E. & Rousseau, Denis-Didier (2024). Matrix profile analysis of Dansgaard-Oeschger events in palaeoclimate time series [Data set]. INESC TEC. https://doi.org/10.25747/T9GX-9729
DOI https://doi.org/10.25747/T9GX-9729
Filstørrelse 8 MB
Format .tex; .csv; .ipynb
Relation 10.5281/zenodo.10527902
Software Spreadsheets, R, python