Scientists Unveil Massive Global Database of Rock Properties Spanning 70 Countries

Every model of the subsurface, whether it is designed to predict how groundwater will move beneath a farming region, how carbon dioxide will behave after injection into a deep saline aquifer, or how heat will migrate around a repository for radioactive waste, depends on one deceptively simple set of inputs: the physical properties of the rocks themselves. Porosity determines how much fluid a rock can store, permeability governs how easily that fluid can move through it, and thermal conductivity controls how heat is transported through the solid framework. Yet for decades, researchers who needed these numbers have faced a frustrating reality. The measurements exist, scattered across thousands of journal papers, government reports and legacy datasets, but they are inconsistent, incompletely documented and often buried in formats that resist modern computational analysis. A team of Australian researchers now believes it has solved that problem on a planetary scale.

In a preprint currently under review for the data-focused journal Earth System Science Data, Heather Anne Sheldon of CSIRO Mineral Resources in Canberra and her colleagues across CSIRO’s energy, environment and information management divisions describe the creation of a Global Petrophysical Database, or GPD, that consolidates more than 112,000 individual rock property measurements drawn from over 600 publications. The compilation spans data from 70 countries and incorporates ten existing databases that had previously been maintained in isolation. Crucially, the database includes both in-situ measurements, taken where the rock actually sits in the subsurface, and laboratory measurements performed on samples brought to the surface, allowing users to compare how properties behave under natural conditions and under controlled experimental ones.

The technical scope of the resource is what sets it apart from earlier attempts. Previous petrophysical databases have tended to be limited in geographic extent, or restricted in the range of properties, rock types or measurement techniques they cover. The GPD deliberately cuts across those boundaries, encompassing igneous, metamorphic and sedimentary rocks alike, and focusing primarily on the properties that control subsurface fluid flow, mass transport and heat transport. Each measurement is accompanied by associated metadata, including rock type, geographic location and depth, the contextual information that transforms a bare number into something a modeller can actually use. Without that context, a permeability value is nearly meaningless, because the same rock type can behave radically differently at two kilometres depth than it does in a surface outcrop.

The practical motivation behind the project is easy to understand for anyone who has attempted a literature synthesis in this field. As the authors note, collating petrophysical property data from the literature is time consuming and is frequently hampered by inconsistent or incomplete metadata. A hydrogeologist modelling aquifer recharge might spend weeks tracking down porosity measurements for a particular sandstone formation, only to find that half the relevant papers omit the depth at which samples were collected or fail to specify the measurement technique. Those omissions matter enormously. Permeability measured on a plug in a laboratory at ambient conditions can differ by orders of magnitude from the effective permeability of the same formation in the ground, where confining stress, temperature and natural fracturing all play their part.

The applications the authors identify span some of the most consequential geoscience challenges of the coming decades. Groundwater management depends on accurate representations of how water moves through aquifers and the low-permeability layers that confine them. Carbon geo-sequestration, the injection of carbon dioxide into deep geological formations as a climate mitigation strategy, requires confident knowledge of reservoir porosity and caprock integrity. Resource exploration, whether for minerals, hydrocarbons or geothermal energy, relies on petrophysical properties to interpret geophysical surveys and to build predictive models of ore-forming fluid flow. And subsurface waste disposal, including the long-term storage of hazardous materials, demands an exceptionally rigorous understanding of how slowly, or quickly, fluids and dissolved species can migrate through host rocks over thousands of years.

What makes the database genuinely powerful for the modern era is its accessibility and its machine-readability. The GPD is publicly available through a persistent digital object identifier hosted by CSIRO, and it is supported by a graphical web interface called the Global Petrophysical Database Explorer, which is designed to facilitate rapid exploration and visualisation of the data. A researcher can query the collection, filter by rock type or property, and visualise distributions without writing a single line of code, while those building numerical models can ingest the underlying data directly. In an age when machine learning is increasingly applied to geoscience problems, a large, well-documented, consistently structured dataset of this kind is exactly the raw material such methods require, and its absence has been a persistent bottleneck.

The scale of the underlying effort should not be understated. Harmonising data from more than 600 publications means confronting a babel of units, measurement standards, rock classification schemes and reporting conventions that have evolved independently across subdisciplines and decades. A permeability reported in millidarcies by a petroleum engineer, one reported in square metres by a hydrogeologist, and a third reported only qualitatively in an old mining report must all be reconciled into a coherent framework. The decision to fold in ten pre-existing databases, rather than starting from scratch, reflects a pragmatic philosophy: the community has already generated an enormous volume of valuable measurements, and the highest-value contribution is to make them findable, comparable and reusable, in line with the open-data principles that journals such as Earth System Science Data were created to champion.

The timing is significant. As governments accelerate investment in carbon capture and storage, geothermal energy and managed aquifer recharge, demand for reliable subsurface property data is rising sharply, and the cost of poor data is measured not just in wasted research time but in flawed decisions about where to inject carbon dioxide or how to site critical infrastructure. A single, globally scoped reference dataset allows modellers to benchmark their assumptions against real measurements from analogous rock types and settings worldwide, rather than defaulting to textbook values that may bear little resemblance to the formations under study. It also exposes gaps: where the database is thin, it points directly to where new field and laboratory campaigns would deliver the greatest scientific return.

The work is currently at the preprint stage, published as a discussion paper with peer review open and ongoing, so the details may evolve as referees and community members weigh in. But the core contribution is already tangible and freely available. For the researchers who built it, the goal was straightforward: to save others the time they themselves had spent hunting for rock property values, and to give the geoscience community a shared foundation for modelling and decision making. If the database achieves the uptake its designers hope, the era of rebuilding the same scattered literature review for every new subsurface project may finally be drawing to a close, replaced by a single, searchable, global record of what we know about the rocks beneath our feet.

Subject of Research: A global petrophysical database compiling rock property measurements for subsurface modelling

Article Title: A global petrophysical database for igneous, metamorphic and sedimentary rocks

Article References: Sheldon, H. A., Dewhurst, D. N., Bekele, E., Raiber, M., Mallants, D., Turnadge, C., & Williams, G. (2026). A global petrophysical database for igneous, metamorphic and sedimentary rocks. https://doi.org/10.5194/essd-2026-556

Image Credits: AI Generated

DOI: 10.5194/essd-2026-556

Keywords: petrophysics, rock properties, porosity, permeability, thermal conductivity, database, groundwater, carbon sequestration, geothermal energy, subsurface modelling, CSIRO, open data