Storage, visualization and simulation using data in Digital Rocks Portal

Imaging such as scattered electron microscopy, magnetic resonance imaging and X-ray micro-tomography have since 1990s introduced 2D/3D datasets of rock microstructure on scales ranging from nanometers to centimeters. The numerical approaches that use such images produce various up-scaled parameters required by subsurface flow and deformation simulators. While these advances have revolutionized our knowledge about grain/pore scale phenomena otherwise impervious to lab measurements, a lack of data-sharing infrastructure among research groups makes it difficult to integrate different length scales required to conduct data-driven explorations.

We have developed a sustainable, open and easy-to-use data management and repository system called Digital Rocks Portal (DRP). DRP is the first of its kind for imaged porous microstructure data (https://www.digitalrocksportal.org). It allows (1) organizing images and related experimental measurements of different porous materials, (2) reference/find them via digital object identifiers (DOIs), and (3) improves their access and analysis to a wider community of engineering or geosciences researchers not necessarily trained in computer science. The data can be uploaded, organized, described, published and visualized remotely on high performance computing cluster while working in your web browser.

We present recent benchmark data sets imaged on multiple scales and stored in the portal. We demonstrate recent portal improvements such as the integration of remote parallel visualization and flow computation workflows with the pore structures currently stored in the repository. We further discuss our perspective on existing challenges for data repositories such as: collecting correct and sufficient metadata, data marketing, and repository sustainability.

DRP is implemented within the reliable, 24/7 maintained High Performance Computing infrastructure supported by the Texas Advanced Computing Center at the University of Texas at Austin. DRP has been developed within the larger set of NSF EarthCube cyberinfrastructure tools (http://www.earthcube.org). This project also supports and aligns with open and Findable, Accessible, Interoperable, and Reusable (FAIR) data principles.