CoolFace
Datasetpublic

Klimkou/GeoGallery

GeoGallery: 3D geological fields Dataset description This dataset consists of geological grids of four different geological sedimentation environments: Barrier Island Tidal Shelf Wave Delta Each grid has 4 different types of fields: facies net-to-gross (NTG) porosity permeability The dataset was created using the framework proposed by K.J. Webber and L.C. van Geuns and then elaborated by N. Tyler and R. J. Finley There are 20004 total geological grids, 5001… See the full description on the dataset page: https://huggingface.co/datasets/Klimkou/GeoGallery.

sourceHugging Facecc-by-nc-4.0updated 2mo agoView on Hugging Face
0likes43downloads
Dataset Card

GeoGallery: 3D geological fields

Dataset description

This dataset consists of geological grids of four different geological sedimentation environments:

  • —Barrier Island
  • —Tidal
  • —Shelf
  • —Wave Delta

Each grid has 4 different types of fields:

  • —facies
  • —net-to-gross (NTG)
  • —porosity
  • —permeability

The dataset was created using the framework proposed by K.J. Webber and L.C. van Geuns and then elaborated by N. Tyler and R. J. Finley

There are 20004 total geological grids, 5001 in for each sedimentation environment. In addition, the dataset has .csv tables of geological statistics:

  • —stats.csv - collector fraction and abruptions
  • —lorenz.csv - Lorenz coefficients
  • —dykstra-parsons.csv - Dykstra-Parsons coefficients
  • —dykstra-parsons-improved.csv - Improved Dykstra-Parsons coefficients
  • —derivatives.csv - Average absolute derivatives of the fields

Dataset structure

The geological cubes are stored in the following way:

{sedimentation}/{field}.npy.zip

where sedimentation is sedimentation environment, field is type of field (ntg, facies, porosity, permeability). The dimension of each field is [5001, 150, 144, 80], where 5001 is total number of grids, 150, 144, 80 is grid resolution

The geological statistics are stored in the following way:

{sedimentation}/{statistics-name}.csv

where sedimentation is sedimentation environment, statistics-name is name of the geological statistics. The dimension of each .csv table is [5001, n], where n is a number of statistics in files.

The custom Croissant metadata is available here.

Loading one 3D geological cube

The repository includes example `load_cube.py` code, which downloads the necessary data and constructs a four-channel cube with sample_id.

Intended use

This geological dataset's intended use is training supervised models and SSL representation models for geological properties prediction.

Limitations and ethical considerations

Limitations

The dataset is limited to four different sedimentation environments, which doesn't cover the full variety of geological objects. It also doesn't cover the hybrid environments.

Ethical considerations

The dataset does not include any personal or other sensitive data.

Citation

When using this dataset, please cite the following paper: