The European Digital Archive of Soil Maps (EuDASM) holds thousands of historical maps as scanned PDF files. Before GIS software can lay a scan over a modern map, the scan must be georeferenced, that is, tied to real-world coordinates. By georefencing, we transform the static EuDASM maps from flat, standalone image files (like JPEGs or PDFs) into spatially accurate, "intelligent" geographic datasets.Georeferencing enables you to overlay these historical soil maps directly onto modern GIS platforms like QGIS or ArcGIS.
In addition, those maps become interoperable with other datasets.

Anyone can help do this with the workflow described below.

What is available?

• A workflow built on QGIS (free software) and a Python script. The script finds the coordinate grid printed on a scanned map, fits the scan to it, checks the fit and writes two GeoTIFFs: one of the whole sheet with its legend, and one of the map face alone. Each result gets a status: PASS, REVIEW, FAIL, TO FILL or NO GRID.
• A download package with the workflow, a short guide (README.txt) and seven worked examples

In a first test of 51 maps (October 2026), 27 became GeoTIFFs (16 PASS, 11 REVIEW). The other 24 need help from a human and spome intervention.

The seven examples listed below show the option to geo-reference a PDF to a TIF file either as a single map or as page (including other info).

- Carte des sols d'Algérie: Alger. Algeria, 1:500,000, base map 1927, completed 1949. PASS.
- Carte des sols d'Algérie: Biskra. Algeria, 1:500,000, base map 1927, completed 1938. PASS.
- Soil Salinity Map, sheet Sidi as Sid. Libya, 1:50,000, 1980. PASS.
- Rabat (topographic map). Morocco, 1:100,000, 1937. PASS.
- Carte de la végétation du Maroc: Rabat-Casablanca. Morocco, 1:200,000, 1955. PASS.
- Carte pédologique de la Tunisie. Tunisia, 1:1,000,000, 1971. PASS.
- Carte des précipitations, average yearly rainfall 1900-1940. Northern Tunisia and north-eastern Algeria, 1:500,000, updated 1948. REVIEW: its grid values could not be confirmed label by label.

The examples cover the three kinds of printed grid the workflow handles: ticks along the frame, kilometre grids, and lines of latitude and longitude. In each, 90% of the printed grid points lie within 0.5 mm, at map scale, of where the georeferencing puts them.

How to start

1. Download the package.
2. Install QGIS from https://qgis.org.
3. Open README.txt in the package. It explains each step.
4. Open the examples in QGIS over an OpenStreetMap background, and check that coasts, rivers and towns line up.
5. Re-create the examples as described in section 2 of README.txt. On Windows this is a double-click; it takes a few minutes and needs no internet connection.
6. Download a map from the EuDASM page: https://esdac.jrc.ec.europa.eu/resource-type/national-soil-maps-eudasm. Put the PDF in the workflow\maps folder, run the launcher "run_georef", type the grid values printed on the map into sheets.csv as described in section 3 of README.txt, and run it again. 
7. Compare the result with a Reference map (including boarders) in QGIS before you share it.
8. Share the results by sending a mail to EC-ESDAC .

Known limits

• Maps without printed coordinates (12 of the 51 test maps) must be georeferenced by hand in QGIS, by matching places on the old map with a modern one.
• Curved, broken or hidden grids, inset maps and sheets pasted together can make the script fail. You then type the grid values, or a few starting points, by hand.
• Most maps do not state their datum, so positions can be off by up to a few hundred metres.
• The automatic checks measure how well a scan fits its own printed grid, not whether the old map is accurate. A person has to compare every result with a modern map.
• The workflow has been tested only on Windows. The launcher for macOS and Linux has not been tested.
• Claude can read the grid values for you through your own paid account. In our test, Claude Code estimated 0.12 to 0.78 USD per map at API list prices. The map images are sent to Anthropic.

You do not need GIS experience to start. Useful tasks at this stage are checking results against modern maps, typing in grid values the script could not confirm, georeferencing maps without printed coordinates, and testing the workflow on macOS or Linux.
 

Figure: Workflow on how to georeference EuDASM maps.

More Information about EuDASM (Proposed citation):

Panagos, P., Jones, A., Bosco, C., Senthil Kumar P.S. European digital archive on soil maps (EuDASM): preserving important soil data for public free access. International Journal of Digital Earth (2011), 4 (5), pp. 434-443. DOI:10.1080/17538947.2011.596580