Laboratory-measured and X-ray CT-derived volumetric composition of a permafrost core

This dataset contains data on the volumetric composition of a permafrost core which has been drilled in a Yedoma upland in northeast Siberia (72.36613 N, 126.27272 E) in September 2017. This dataset supplements a research article to be submitted to the scientific journal The Cryosphere. It contains the following files: volumetric_contents_sampleRes_lab+CT.csv Contains the volumetric contents of total ice, organic, and mineral measured in the laboratory at AWI Potsdam at a coarse resolution. It further contains the volumetric contents of gas, excess ice, and two sediment phases (A,B) derived from a CT scan at UFZ Halle, downsampled to the resolution of the laboratory samples. volumetric_contents_highRes_CT.csv Contains the volumetric contents of gas, excess ice, and two sediment phases (A,B) derived from a CT scan at UFZ Halle at the original resolution of 50µm. regression analysis_paper.py This pyhton script uses the above listed input files to perform and evaluate a regression analysis of the CT data against the laboratory data. The regression result is the composition of the CT-derived sediment phases (A,B) in terms of pore ice, organic, and mineral. The script furthermore computes evaluation metrics of the lab-CT comparison, and computes volumetric contents of pore ice, total ice, organic, and mineral at the high resolution of the original CT data. volumetric_contents_sampleRes_all.csv This file can be reproduced by the files listed above and contains, in addition to the data contained in volumetric_contents_sampleRes_lab+CT.csv, the volumetric contents of pore ice, total ice, mineral, and organic as predicted by the regression model at the same (coarse) resolution as the laboratory samples. volumetric_contents_highRes_all.csv This file can be reproduced by the files listed above and contains, in addition to the data contained in volumetric_contents_highRes_CT.csv, the volumetric contents of pore ice, total ice, mineral, and organic as predicted by the regression model at the same (high) resolution as the original CT data. More details can be found in the article describing the study.