Research Data
Data Publication
Data for Generative AI Modeling of Serpentinte Dehydration Veins in Subduction Zones
Austin Arias
Utrecht University
(2026)
Descriptions
We used hand specimens of Erro-Tobbio meta-serpentinite (ET-MS) and outcrop surface images of olivine dehydration veins to train generative models to generate unique and varied dehydration vein network structures. These models coupled with thermodynamic calculations allowed us to predict maximum permeabilities of serpentinite undergoing antigorite and brucite dehydration. The data is organized into either a shared "Data" folder or a "Figures" folder. Data used to plot multiple figures is provided in the top-level "Data" folder. This includes 2 folders, one for EPMA data describing mineral compositions of ET-MS and another for the results of permeability calculations for the ET-MS generated volumes. Within the top level "Figures" folder there are 11 folders specific to each figure of the main publication. Within each, the data required to recreate each figure as well as scripts used to plot each figure are provided.
Keywords
MSL enriched keywords
minerals
silicate minerals
nesosilicates
olivine
metamorphic rock
serpentinite
Measured property
permeability
Apparatus
X-ray tomography
Technique
imaging (3D)
computed tomography (CT)
equipment
electron probe micro-analyzer
microchemical analysis
electron probe micro analyser
tectonic plate boundary
convergent tectonic plate boundary
subduction
Analyzed feature
deformation microstructure
pressure solution microstructure
vein
oxide mineral
brucite
phyllosilicates
serpentine
antigorite
MSL vocabulary keywords corresponding to originally assigned keywords
olivine
serpentinite
permeability
X-ray tomography
computed tomography (CT)
electron probe micro-analyzer
electron probe micro analyser
Originally assigned keywords
FOS: Earth and related environmental sciences
olivine
serpentinite
permeability
porous media
artificial intelligence (AI)
generative adversarial network (GAN)
thermodynamics
X-ray tomography
scanning electron microscopy
electron microprobe
Metadata
MSL enriched sub domains
rock and melt physics
microscopy and tomography
geochemistry
Resource Type
Source publisher
| Utrecht University |
Creators
| Austin Arias |
| Personal |
| https://orcid.org/0000-0002-6575-4191 |
| Utrecht University |
Citation
Arias, A. (2026). Data for Generative AI Modeling of Serpentinte Dehydration Veins in Subduction Zones [Dataset]. Utrecht University. https://doi.org/10.24416/UU01-MY6CEC
References
Dates
| Issued | 2026-06-08T09:58:02 |
| Updated | 2026-06-08T09:58:16 |
Language
en
Funding References
| Funder Name | NWO |
Rights
| Name | Open - freely retrievable |
| URI | info:eu-repo/semantics/openAccess |
| Name | Creative Commons Attribution 4.0 International |
| URI | https://creativecommons.org/licenses/by/4.0/legalcode |
| Identifier | cc-by-4.0 |
| Identifier Scheme | SPDX |
| Scheme URI | https://spdx.org/licenses/ |
Locations
Geo location(s)
Erro-Tobbio, Voltri Massif, Liguria, Italy