Seismic and Electrical Resistivity Tomography of Lagoon Pingo, Svalbard in 2019

This mixed dataset contains raw and processed seismic and electrical resistivity tomography (ERT) data from Lagoon Pingo, Svalbard. Lagoon Pingo is a coastal open-system pingo with year-round methane release located in a glacio-isostatically uplifting fjord-valley. This dataset arises from our geophysical investigations of Lagoon Pingo, which aims to characterise and understand how pingo structure and composition facilitates methane release through continuous permafrost.

Two active source seismic profiles were collected using two Geometrics Geode 24-channel seismographs between 21/08/2019 and 29/08/2019, which together cover the longest lateral extent of the open-system pingo. Seismic profiles consisted of 48 electrodes spaced at 4 m intervals, with shots taken at 4 m intervals between electrodes and up to 100 m off-end using a sledgehammer on a metal baseplate. This is complemented by 5 ERT profiles taken between 05/04/2019 and 10/04/2019, collected using an ABEM SAS1000 Terrameter, with short profiles containing 21 electrodes covering the width of the pingo, and longer profiles of 53 electrodes covering the longest lateral extent. Data were collected using a Wenner configuration and automatic selection of input current.

Dataset Contents

This dataset contains the following data:

• Raw seismic shots taken from Lagoon Pingo in SEG-2 format
• ReflexW files for the seismic tomography and wavefront inversion derived from the first onset of these single shots
• The processed common offset gather (a.k.a. "supergather") derived from these shots
• Raw, unfiltered ERT data in ASCII format
• Inverted ERT profiles, as obtained following inversion in ResIPy, with files in ASCII and .vtk format

File Structure

• Seismics
  • Raw
    • S1 – 71 folders with one for each shotpoint, each containing 10 .SG2 SEG-2 files (710 files total)
    • S2 – 71 folders with one for each shotpoint, mostly containing 10 .SG2 SEG-2 files (some folders have a slightly different number of shots due to miscounting when taking the data; 718 files total)
    • A separate readme file (markdown format: .md), which provides information regarding filenames and their relative locations.
  • Processed
    • Seismic Refraction
      • Tomography – 1 folder (containing 5 files) and 7 separate files)
      • Traveltimes – 1 ASCII file.
      • Wavefront Inversion – 4 ASCII files
    • Supergather – 1 ReflexW processed file
• ERT
  • Raw – each folder (5 total) pertains to one profile, with each containing the following (20 files in 5 folders):
    • 2 csv files highlighting electrode and finer resolution (2.5 m) topography
    • 2 ASCII files with ERT data – normal and reciprocal for each profile.
  • Inverted – each folder (5 total) pertains to one profile, with each containing the following (5 x 11 files; 55 total files):
    • 2 .vtk files which allow visualisation of the inversion and relative location of the electrodes respectively in an open-source program such as Paraview
    • An ASCII file called electrodes.dat, which provides the line topography used in the profile
    • 3 ASCII files, called “f001_err.dat”, “f001_res.dat” and “f001_sen.dat”, which provide the error, resistivity and sensitivity data of the final inversion respectively
    • An ASCII file called mesh.dat, which provides the mesh constraints used in the final inversion
    • An ASCII file called protocol.dat, which provides the measurement configuration for data used in the inversion
    • An ASCII file called R2.in, which provides the parameterisation used for the inversion
    • An ASCII file called R2.out, which provides the output log from the inversion
    • An ASCII file called res0.dat, which provides the parsed data input for the inversion

The total size of the repository is approximately 1.12 GB.

Methodology

Seismic data were acquired using two Geometrics Geode 24 Channel seismographs in profiles that consisted of 48 100 Hz geophones at 4 m spacings. Zero offset shots were gathered using a sledgehammer source on a metal baseplate at 4 m intervals between geophones and up to 10 m off-end, with additional shots taken at 10 m up to 100 m off the end of the profile.

Seismic data were processed in REFLEXW version 9.5 (Sandmeier, 2020). First, shot geometry was adjusted according to GPS points taken using a Garmin eTrex handheld GPS at c. 5 m accuracy, and topography obtained from a 2010 summer DEM (Norwegian Polar Institute, 2014). Second, traces that were exceptionally noisy or occurred due to dead electrodes were removed. The common offset gather was produced using 4 m spaced shots through the following steps:

1. Adjustment of traceheaders to use relative geometry, and the merging of all shotgathers from both lines into a single profile;
2. A resorting of shots according to offset between source and receiver;
3. Splitting the merged file according to common offsets between traces;
4. Stacking the split profiles containing common offsets, to provide a singular trace for each offset value;
5. A recombination of all common offset traces; and a resorting of this according to the traceheader to provide a common offset gather;
6. And lastly, divergence compensation for display.

For seismic refraction:

1. First onsets were obtained for each shotpoint within the 4 m spaced shots;
2. Traveltime analysis was conducted as per standard REFLEXW procedure (Sandmeier, 2020);
3. First arrivals were assigned, and the wavefront inversion (which uses a finite difference calculation) and tomography (which uses the Simultaneous Inversion Reconstruction Algorithm) modules of the REFLEXW software were initiated.

The tomographic inversion was parameterised following initial seismic refraction analysis of the common offset gather as below:

Space increment: 0.5

Maximum number of iterations: 20

Maximum time: 10800 seconds

Threshold: 0.001

Model change A: 1

Model change B: 0.1

Convergence search: 5

Max def. change (%): 200

Def. data-variance: 0.01

Maximum beam width: 10

Start curved ray: 1

Average x smoothing: 2

Average z smoothing: 0

Minimum velocity: 1

Maximum velocity: 4000

Electrical resistivity tomography data were acquired using an ABEM SAS1000 Terrameter, through profiles of two different lengths: shorter transects consisting of 21 electrodes with 5 m spacing, and longer profiles of 53 electrodes with 5 m spacing. Wenner array configurations were used owing to variable and uncertain field site conditions, and input current was automatically determined by the instrument. To counteract electrode contact issues in frozen conditions, a saline solution was poured onto each electrode before measurements were made. The location of selected electrodes were taken with a Garmin eTrex handheld GPS at c. 5 m accuracy, and topography was obtained by interpolating these points to provide a location for each electrode, and deriving elevation from a 2020 Structure-from-Motion DEM (Hann and Dachauer, 2020).

The inversion of data was completed using ResIPy version 3.3 (Blanchy et al., 2020). Within ResIPy, the following process was used:

1. Import of data and parsing from ASCII format to that recognisable by ResIPy, and import of the calculated profile topography;
2. Filtering of data, using a 10% threshold between normal and reciprocal measurements;
3. Selection and fitting of an error model, with linear models used for profiles composed of 21 electrodes and power-law models for those composed of 53 electrodes.
4. The construction of a quadrilateral mesh within the software, using a horizontal resolution of 2 cells per electrode and default parameters elsewhere;
5. Inversion using the following parameterisation:

Inversion type: Regularized Inversion with Quadratic Filtering [2]

Target decrease: 0

Data type: Logarithmic [1]

Regularization mode: Normal [0]

Value for tolerance: 1.0

Maximum number of iterations: 30

a_wgt, b_wgt: 0, 0

Minimum apparent resistivity: -10e10

Maximum apparent resistivity: 10e10

Maximum beam width: 10

Flux Type: 3D

Value for res_matrix: Sensitivity matrix [1]

Patch size x, y: 1, 1

Update the weights: Recommended [2]

Value for alpha_aniso: 1

Where inverted data points occurred outside of a ±3% threshold, these points were removed and data subsequently reinverted using identical parameters.

Spatial Extent

For the seismic dataset, topographic information is included within the fileheader for each shot. For the ERT dataset, the location of each electrode is presented within the electrodes topography csv for each profile. For both datasets, co-ordinates are displayed in UTM 33N (ESPG: 25833).