Dataset DOI: 10.5061/dryad.5x69p8dj9

Description of the data and file structure

This Zip file contains CT data of sediment cores from the Hamilton lowlands, New Zealand. CT stacks of a total of 60 sediment cores from 18 different lakes are included. The following data processing steps were undertaken.

CT imaging was performed using a medical CT scanner on 60 whole-round cores prior to opening (i.e., all lake cores except HL-A2 and HL-A3 from Rotoroa/Hamilton Lake, which were not available for CT imaging) to visualize and unambiguously quantify tephra layers and soft-sediment deformation structures. The whole-round core was positioned in the center of the CT table, and preliminary scan results of the tephra layers and tephra seismites could be observed in the CT control room (fig. S1AB).

A CT stack (sequence of cross-sectional CT images) was generated for each core, with each CT image comprising voxels with dimensions of 0.625 mm depth and 0.684 mm width and height. The voxel values represent the radiodensity in Hounsfield units HU. An HU value of -1000 represents the radiodensity of air, and an HU value of represents the radiodensity of distilled water at standard temperature and pressure. CT image processing was performed using the public domain image processing software package ImageJ and its portable distribution Fiji (version 1.53t) and included the following steps (62):

(a) Reduction of CT data: The raw CT stack initially consists of the core barrel (filled with sediment, water, and tephra) located longitudinally in the center of the CT table, both surrounded by air. To reduce computation time, the CT stack was roughly cropped to the data representing a cuboid around the core barrel using Fiji's rectangular selection tool.

(b) Combination of CT stacks: The majority of the sediment cores were short enough to be acquired in a single CT scan, but some of the longer cores extended beyond the measurement area of the medical CT scanner. For these cores, distinct layers were identified in the overview core scan, and CT imaging was then performed once from each core end, allowing the distinct layers to be recorded in both scans, but in different directions. The reverse-direction CT scan was rotated 180 degrees (using Fijis rotation tool) and then combined with the first CT scan using the distinct layers as a reference.

(c) Core rotation and realignment: Cores were sometimes not perfectly aligned longitudinally on the CT table. In addition, longer cores could not be placed flat on the CT table (due to a shelf space at the base of the CT table) and instead were slightly tilted upward at one end of the CT table, causing the core to be slightly rotated or bent upward, or both. The CT stacks were reoriented using Fiji's rotation tool. The curvature of the cores was removed using the open-source Fiji stack alignment plugin.

(d) Removal of voxels outside the sediment core: After the core was properly rotated and aligned, the data representing the core barrel, CT table, and air were removed using Fiji's circular selection tool. The remaining CT stack represented only the inside of the core barrel (i.e., sediment, water, and tephra) in a cylindrical shape.

(e) Contrast and brightness enhancement was applied to highlight the different sedimentary units in the core, mainly the organic lake sediment, tephra layers, and soft-sediment deformation structures. The raw CT images that make up a CT stack use a standard linear transfer function between radiodensity and a grey scale value. Typically, black represents the lowest HU value (air) and white represents the highest HU value in the core. The sedimentary units in the core have relatively similar densities compared to the surrounding air, so the core appears white in the raw CT images (i.e., very little change in pixel values within the core) (first panel of fig. S1C). The contrast of CT images was enhanced using Fiji's enhance contrast tool. Voxel values were normalized to the highest value of the entire CT stack, with 0.35% of all voxels saturated (i.e., assigned a maximum value) (second panel of fig. S1C). A custom transfer function (mpl-viridis) was applied to the CT images to better highlight small density changes within tephra layers (third panel of fig. S1C). The custom transfer function assigns dark blue/purple to HUmin, green to 0.5HUmax, and light yellow to HUmax.

(f) Substacking: Stratigraphic units (i.e., tephra layers of known origin and age, organic lake sediment, pre-lake alluvial sediment) were identified for each sediment core using information from the core description. Once the stratigraphic framework was established, substacks were created for the relevant tephra layers (Rk, Rr, Oa-8, Wh, Mm, Op, Rm, Ma, Tu, Tn, UnitQ, Wo, Mp, Tp). In total, 346 tephra layers were visualized (fig. S2).

(g) Visualization of CT substacks: Each substack, representing one or more tephra layers, was visualized in three ways (fig. S1D): (1) Longitudinal CT slices were generated through the center of the core. Typically, three longitudinal CT slices were generated at angles of 0, 120, and 240 degrees (where 0 degrees was defined as parallel to the cut core surface) to facilitate a systematic analysis of the vertical and horizontal extent of tephra layers and tephra seismites. Longitudinal CT slices were best suited to analyze the thickness of tephra layers as well as the maximum vertical length of soft-sediment deformation structures. (2) A series of cross-sectional CT slices was generated for each CT stack containing tephra seismites to better visualize the changes in thickness and morphology of individual soft-sediment deformation structures with depth. Cross-sectional CT slices were used to assess the thickness and interconnectivity of individual soft-sediment deformation structures. (3) Three-dimensional CT volumes were generated from each CT stack using Fiji's volume viewer plugin. The organic lake sediments above and below the tephra layers were digitally removed by applying an -function based on X-ray attenuation, which was significantly different between the two constituents (i.e., organic lake sediments versus tephra) and allowed images to be exported showing only the tephra deposit. Standardized images were generated for each CT volume, including three images orthogonal to the core taken at angles of 0, 120, and 240 degrees and two oblique images (one from above and one from below). CT volume images, in which the low-density organic lake sediment had been filtered out, provided the best tool for studying the three-dimensional morphology of the tephra seismite and tracking individual soft-sediment deformation structures. A compilation of all CT substack visualizations is provided in data S1.

Files and variables

File: CT_data.zip

Description: Zip folder contains 19 folders, one for each lake for which CT scans were obtained. Each of the 19 folders on hierarchy level 1 contains a number of subfolders (hierarchy level 2) that show individual CT scans of cores. Hierarchy level 3 shows screenshots of the processed CT data for each scan. These screenshots show the CT data from different perspectives (e.g., longitudinal and cross-sectional). Some of the Scans contain videos of CT scans. Hierarchy level 4 (always called IS) contains the processed CT slices as .tiff