the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Oct 2026
A global petrophysical database for igneous, metamorphic and sedimentary rocks
Abstract. Petrophysical properties such as porosity, permeability and thermal conductivity are required to support modelling and data interpretation in many geoscientific fields, including groundwater management, carbon geo-sequestration, resource exploration, and subsurface waste disposal. Collation of petrophysical property data from the literature is time consuming, often hampered by inconsistent or incomplete metadata. Existing petrophysical property databases are limited in their geographic extent, or in the range of included properties, rock types or measurement techniques. Motivated by these challenges, a Global Petrophysical Database (GPD) has been developed (Sheldon et al., 2026), incorporating data from more than 600 publications including 10 existing databases and covering 70 countries. Both in-situ and laboratory measurements are included. The database focuses primarily on properties that control subsurface fluid flow, mass transport and heat transport. The GPD is publicly accessible at https://doi.org/10.25919/e8f9-n090 (Sheldon et al., 2026), supported by a graphical web interface, the Global Petrophysical Database Explorer, to facilitate rapid exploration and visualization of the data.
- Preprint
(1055 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 26 Nov 2026)
- CC1: 'Comment on essd-2026-556', Andrew Cavanagh, 23 Oct 2026 reply
Data sets
Global Petrophysical Database Sheldon, H.A., Bekele, E., Dewhurst, D.N., Mallants, D., Raiber, M., Turnadge, C. https://doi.org/10.25919/e8f9-n090
Model code and software
Global Petrophysical Database Explorer Sheldon, H.A., Williams, G., Mallants, D. http://hdl.handle.net/102.100.100/600171?index=1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 409 | 98 | 28 | 535 | 33 | 28 |
- HTML: 409
- PDF: 98
- XML: 28
- Total: 535
- BibTeX: 33
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dear Dr Sheldon,
First, a big thank you to you and your co-authors for openly releasing the Global Petrophysical Database! It is a very valuable resource, and we are using v6 (doi:10.25919/e8f9-n090) as the mudrock baseline for work on caprock permeability and basin-scale COâ‚‚ storage.
While checking our analysis against the original sources, we compared the ‘Yang2007’ records (rows 81103–81570) cell by cell with Yang and Aplin (2007, JGR 112, B03206, Tables 1–3). We noted a small number of discrepancies.
We report the results below to help with the next version. Most of the transcription is accurate. All Table 1 values for clay fraction, porosity, grain density, specific surface area and TOC match the paper exactly.
(a) Row 81404 (Yang2007_00019, paper sample 20, vertical, 15 MPa) gives 2.1E-19 m². Table 3 gives 2.1E-20 m², which also fits the monotonic decline from 2.6E-20 at 10 MPa to 1.7E-20 at 20 MPa.
(b) Row 81482 (Yang2007_00024, paper sample 25, vertical, 45 MPa) gives 8.0E-23 m². Footnote c of Table 3 assigns 8.0E-23 m² to sample 18 and 1.2E-22 m² to sample 25. Sample 25’s 45 MPa value appears to have been replaced by sample 18’s value, and there is no 45 MPa row for sample 18 (Yang2007_00017). As entered, sample 25 also becomes non-monotonic (8.0E-23 at 45 MPa, then 9.0E-23 at 50 MPa).
(c) Footnote b of Table 3 reports three unloading permeabilities for sample 7 (horizontal): 9E-21, 1.2E-20 and 1.9E-20 m² at 40, 25 and 10 MPa. They are not in the database (Yang2007_00006, rows 81205–81217). Leaving out unloading data is reasonable, but a note in Remarks would tell users the values exist.
(d) Effective stress. The remark ‘Estimated in-situ effective stress X MPa’ carries the paper’s ‘Maximum effective stress’ (Table 1), not its ‘In situ effective stress’. For sample 1 (row 81103 onwards) the remark gives 28 MPa, but the paper reports no in-situ value. For sample 3 the remark gives 23 MPa, against 24.6 MPa in situ and 23 MPa maximum in the paper. Users who select data at in-situ conditions may be misled.
(e) Depth datum. ‘Depth Min (m)’ takes the paper’s ‘Burial depth’ where one is reported, and ‘Depth’ otherwise. For example, sample 2 has 1860.9 m (burial) against 2047.9 m (depth), while sample 1 has 3695.1 m (depth; no burial value given). The column therefore mixes two datums within one reference.
(f) Lithology of sample 1 (Yang2007_00000). It is entered as ‘mudstone’, but it contains 65.4% sand, and Yang and Aplin classify it as a muddy sandstone (Figs 4–5; text paragraph 27).
(g) Duplicate grain densities. Paper samples with both horizontal and vertical measurements have two grain-density rows. For samples 18 and 30 the two rows differ (rows 81344 and 81357: 2812 and 2810 kg/m³; rows 81545 and 81559: 2782 and 2780 kg/m³). Table 1 gives 2812 and 2782.
(h) Dry density. ‘Density - dry’ values (for example 2330.084 kg/m³ for sample 1) are not tabulated in the paper. They equal grain density × (1 – porosity), but they are labelled Measurement type ‘lab’. ‘Lab (computed)’, or a remark, would make the derivation clear.
(i) Quality. All Yang2007 permeability rows are ranked D, with preservation ‘Dry’. The paper (paragraph 6) says some samples ‘had been recently cored and were wet, but many had dried as a result of storage’. A single ranking may be the only practical choice, but a remark would help.
(j) Version history versus content. The v6 version history says ‘Populated Quality column for all lab permeability measurements on shale/mudstone/claystone’. We count 125 lab permeability or hydraulic-conductivity rows with Lithology simplified = shale, mudstone or claystone that still have a blank Quality (104 of them from Neuzil1986).
(k) Representative values for ranges. Of the 1,453 permeability or hydraulic-conductivity rows that report both Min and Max, 165 have a Value equal to the arithmetic mean of the two (for example row 666, Armitage2016). Where this comes from the compilation rather than the source, a geometric mean, or a note, would suit log-normally distributed properties better.
(l) Earlier versions. Some catalogue mirrors (for example data.gov.au, harvested from CSIRO DAP) still point to an earlier version under doi:10.25919/kfvt-2197. A cross-reference between the two DOIs would help users confirm they have v6.
I hope the above information helps. Thank you again for making this database available.
Best wishes,
Andrew
Dr Andrew Cavanagh
SCCS, Edinburgh Climate Change Institute
School of GeoSciences, University of Edinburgh