The data

The data used for Nuclear Cartographies was organized into four categories, all pertaining to France:

On this page, you can read about how the data was found, collected, sorted, and cleaned for the project.

Atomic tests:

The dataset for atomic weapons was collected on the community-led website “overpass turbo” — a tool which pools data from different sources for OpenStreetMap.

Since the data entered on overpass is multiple in origin (i.e. has multiple contributors) and is entered following broad categorizations on a CSV format (for example, the dataset for atomic tests contained categories such as “crater dimension” or “wave amplitude” that were often left empty or simply did not apply), this “raw” data was then checked and completed individually using Bruno Barrillot’s inquiry L’héritage de la bombe : Sahara, Polynésie (1960-2002) to make sure that every entry was correct, then “cleaned” using a simplification procedure that eliminated missing, vague, or redundant information before inputing it on the map.

Nuclear reactors:

The raw data for nuclear reactors was pooled from a database created by Logan Byers, Johannes Friedrich, Roman Hennig, Aaron Kressing, Xinyue Li, Colin McCormick and Laura Malaguzzi Valeri, and published online by World Resources Institute. The data was then checked, cleaned, and completed by using information provided by the French electricity company EDF.

Nuclear waste repositories:

The data for nuclear waste was collected online via the French agency ANDRA (Agence Nationale pour la gestion des Déchets Radioactifs). This governmental website catalogues different kinds of broadly defined wastes from the nuclear industry and generated by various organizations, ranging from military defense usage, to the uranium mining system, to that of medical facilities. As such, the nature of the waste varies greatly depending on the repository site.

The kinds of waste were sorted by ANDRA into 7 families, which were left untouched for the map:

  • High Activity (Haute Activité): a waste whose radioactivity is extremely high, reaching up to several billions of Becquerels per gram (read more about units of measuring radiation here).
  • Medium Activity with Long Life (Moyenne Activité à Vie Longue): a waste whose radioactivity is moderate (from a few millions to a billion Becquerels per gram) but whose half-life is long (superior to 31 years).
  • Low Activity with Long Life (Faible Activité à Vie Longue): a waste whose radioactivity is weak (from a few hundreds to a million Becquerels per gram) but whose half-life is long (superior to 31 years).
  • Low to Moderate Activity with a Short Life (Faible à Moyenne Activité à Vie Courte): a waste whose radioactivity is weak to moderate (from a few hundreds to a billion becquerels per gram) and whose half-life is short (inferior to 31 years).
  • Very Low Activity (Très Faible Activité): a waste whose radioactivity is inferior to 100 becquerels per gram.
  • Very Short Life (Vie Très Courte): a waste whose half-life is very short (inferior to 100 days).
  • Others (Autres): a highly variable category of waste which was often (though not always) assigned a subcategory. This family of waste, by far the most statistically common, varied greatly. Some of the repository sites, for instance, contained waste from building materials, aviation, or even buttons painted in fluorescent paint that contained a radioactive material. Others, like the sites of Moruroa and Fangataufa, contain the wastes of hundreds of atomic bombs that were not always carefully sealed. The volume, too, varied greatly (when it was indicated), from several buttons to thousands of cube meters of highly radioactive material.

For the purpose of cartographic economy, waste from medical facilities and universities were not included, primarily because the entries were almost as numerous as there are hospitals and universities with a physics department, which would not give an accurate understanding of waste sites proper; second, because the vast majority of them were catalogued as VTC (Very Short Life) or TFA (Very Low Activity) and are therefore on a very short turnover, sub-labeled DGD (Déchets Gérés en Décroissance, that is, a type of waste managed on site as they lose their radioactivity until they can be re-used). In other words, the nuclear waste categories favored for this project were either highly radioactive waste, waste that could not be reconditioned, or both.

The nomenclature, the sorting system, and the information pertaining to the waste recorded on the ANDRA website, however, were contingent to say the least: although there was some indication of volume when a description of the nature of waste was provided, total volume as well as radioactivity (in Becquerels) were missing in the vast majority of cases. Descriptions of waste, too, were truncated from the website and could not be reliably read. For nuclear waste originating from atomic tests, this lack of transparency is highly problematic if unsurprising.

Finally, ANDRA published a report on submerged nuclear waste — an internationally coordinated practice which France partook in on two occasions: once in 1967 and once in 1969. The report, however, did not provide geolocations for these sites (only vague descriptions). To find these sites, it was therefore necessary to cross-reference these information with the IAEA’s own report on “Inventory of radioactive waste disposals at sea” in order to collect the geospatial data.

Uranium mines:

The French history of mining uranium is long and complex, involving multiple organizations:

  • The CEA (Commissariat à l’Énergie Atomique, or the Atomic Energy Commission) which was in charge of both prospecting and extracting the ore until the 1970s
  • Cogema (Compagnie générale des matières atomiques, or General Company of Atomic Matters) which took over mining activities from the CEA in 1976.
  • AREVA, created through a merger of different French nuclear industrial groups in 2001, to oversee many of activities of the nuclear cycle, including (but not limited to) uranium mining.
  • Orano, born in 2017 out of a scission from AREVA, whose sole industrial activity is that of uranium.

French uranium mines, for the purpose of Nuclear Cartographies, were sorted into two types:

  • Mines in the metropole (which, as of 2001, are no longer operational);
  • Mines outside of the metropolitan borders but exploited by a French company (either the CEA, Cogema, AREVA, or Orano).

The data for metropolitan mines was directly collected from “Projet Mimausa” — a database and cartographic projection of every uranium mining sites across metropolitan France conducted by the governmental agency IRSN (Institut de Radioprotection et de Sûreté Nucléaire, or the Radioprotection and Nuclear Safety Institute).

For data on uranium mines laying outside of French borders, the process was more complicated, not only because there is no publicly accessible database, but also because many of the mines exploited by the CEA (the French equivalent of the Atomic Energy Commission) in the 1940s and 1950s, particularly in Madagascar, have long been abandoned and dug over. Their records, too, were not transferred to the following organizations that took charge of mining (Cogema, AREVA, Orano). To find the names of the sites was therefore the first challenge; finding their accurate geolocation was another.

To identify and accurately place some of the mining sites in Madagascar and Gabon (though it is highly probable that dozens of other sites are still missing), I used the following documents:

Other mining quarries (post-1970s) were identified through the data provided on Orano’s website and cross-referenced using Google Map before being manually entered on the data sheet.