Reggane: 17 tests, 11 after independence, 73 requests — 1 granted

Nadia Cherifi — Alger, Algérie — 2026-09-11

Note on method, and it is essential here. This dossier concerns a subject on which many false figures circulate — on both sides. We therefore worked exclusively from institutional documents: the reports of the International Atomic Energy Agency, of UNSCEAR (the United Nations scientific committee on the effects of radiation), of the French Parliamentary Office for the Evaluation of Scientific and Technological Choices (OPECST), of the French National Assembly and Senate, of the Committee for the Compensation of Victims of Nuclear Tests (CIVEN), of the Council of State and of the Paris Administrative Court of Appeal. We systematically state what France acknowledges, what it refuses, and what no one has ever established. We set aside the figures we were unable to verify ourselves, even when they serve us. A dossier on Reggane that exaggerates is a dossier lost.

☢️ Seventeen tests, eleven after independence

Let us begin with the facts, because they are heavy enough that there is no need to add to them.

France carried out seventeen nuclear tests on Algerian soil: four atmospheric at Reggane, thirteen underground at In Ekker.

These figures are corroborated by UNSCEAR (2024 report, annex B), by the International Atomic Energy Agency (1999 mission, report published in 2005), by the Committee for the Compensation of Victims of Nuclear Tests, and by two reports of the French National Assembly — No. 1558 of June 10, 2025 and No. 2364 of January 21, 2026.

And here is the fact everyone knows but few people measure: eleven of those seventeen tests took place AFTER Algeria's independence.

It is the French National Assembly that writes it, in black and white, in its report No. 1558: “of the seventeen tests carried out by France in Algeria, eleven were after independence”.

Independence was proclaimed on July 3, 1962. The last test took place on February 16, 1966. For nearly four years, a sovereign state was the theater of nuclear experiments conducted by another state.

And there is one case that deserves a pause: the Béryl test, on May 1, 1962. It falls between the Évian Accords, signed on March 18, 1962, and the proclamation of independence. That is, in that legal in-between where Algeria was no longer quite French and not yet independent.

⚠️ The question no one has settled: the Évian Accords

We searched, text in hand, for what the Évian Accords say about nuclear tests.

Answer: nothing. The word “nuclear”, the word “test”, the names “Reggane” and “In Ekker” do not appear in ANY of the published texts. We checked two different full reproductions.

The only published textual basis is a general formula: “Algeria also grants to France the use of certain airfields, lands, sites and military installations necessary to it.”

On the other hand — and this is an assertion of the National Assembly, not ours — the two parliamentary reports of 2025 and 2026 indicate that annexes not published at the time specified that France obtained the right to use the sites of Reggane and In Ekker for nuclear tests for five years after independence.

We give both positions, because both exist:

• the French Parliamentary Office writes that “in accordance with the Évian Accords, the sites […] were handed back to the Algerian authorities in the course of 1967, after […] the cleaning and sealing of the galleries”;

• a Princeton researcher, in a 2025 paper, writes that the Évian Accords “did not oblige France to dismantle and monitor the test sites”.

No judge has ever settled the scope of that annex. We will not settle it in his place.

💥 Yields: three official sources, three sets of figures

This is the most contested point of the dossier, and it is also the one where caution is most necessary. Three official institutions give three different versions.

What this concretely means: the total yield of the Algerian tests has never been established with certainty. We give the institutional ranges and we refuse to settle it.

What we will not write: the precise values widely circulating on the internet — 70 kt for Gerboise bleue, 127 kt for Saphir, a total of 285 kt — come from no institutional document. We looked for them in the reports of the IAEA, the Senate and UNSCEAR: they are not there.

🩺 The Béryl accident, and the documented doses

On May 1, 1962, at In Ekker, the underground test Béryl went wrong. The gallery plug gave way, and part of the radioactivity escaped into the atmosphere.

The official figures, as they appear in the report of the French Parliamentary Office:

And here are the cumulative doses of the personnel, established by the French monitoring commission (2007 report):

One fact we must state, because it is true: nine of the thirteen underground tests were fully contained. Béryl and Améthyste were partially contained; Rubis and Jade were not fully contained (IAEA 2005 report). We will not claim that everything leaked: that would be false and immediately refutable.

🔬 The IAEA report: what it really says — and what it does not say

It is the most cited document in this dossier, on both sides. It must therefore be read precisely.

What it is: a mission carried out in Algeria from 18 to November 26, 1999, at the request of the Algerian government. Three sites visited, 76 dose-rate measurements, 25 samples analyzed in Austria. The report was published in 2005.

Its conclusion: the clean-up of the test sites “is not necessary” — provided that local economic conditions do not change.

And now, what the report says about itself. Here are its own words:

“preliminary assessment” […] “a limited and minimal level of sampling” […] “the measurement data provided here would be inadequate” […] “this low ratio between the number of samples and the total area […] restricts the general validity of the conclusions” […] “the dose estimates […] are highly uncertain”.

And above all — this is the decisive point of the dossier — the IAEA report contains NO health data. Not one study, not one case, not one rate. In 71 pages, nothing on the health of the populations. The reference document on the Algerian sites has never assessed anyone's health.

Two admissions of dependence on France also appear in the report:

“The mission was able to accomplish more than foreseen as a consequence of the detailed information provided by France”; “the contribution of the French expert within the team […] was invaluable”.

And a final element, which dates from before: in 2000, the reference scientific compilation of the United Nations stated in black and white: “There is no information on local exposures following these tests.”

🧬 Health: what is established, what is probable, what is contested

This is where we must be most honest, because this is where the debate is most impassioned — and most poorly documented.

WHAT IS ESTABLISHED

• The documented individual doses, as listed above.

• Residual contamination still measurable in 1999 at the ground zero points of Gerboise bleue and Gerboise blanche at Reggane, and in the lavas of tunnel E2 at In Ekker (cesium 137, strontium 90, plutonium).

• The official French position: it is the report of the monitoring commission (CSSEN) of July 3, 2007 that concludes that “no new risk of radiation-induced cancers has been demonstrated”. Note: there is no IRSN report on the Algerian sites.

WHAT IS PROBABLE

• A link deemed “likely” in 6 of 15 cases by INSERM forensic expert appraisals, commissioned by a French investigating judge, delivered in 2009-2010 and revealed in July 2012. Crucial precision: those appraisals concerned French veterans, Algeria and French Polynesia combined — not Algerian civilians. The expert himself acknowledges: “one will never be able to assert such or such causality with certainty”, and “there remains information classified as defense secret”.

• The local cancer registries. The Adrar registry recorded 183 new cases in 2019, 181 in 2018, 163 in 2017, 102 in 2016. Caution: these data are described as “preliminary” and “provisional” by their own coordinator. In Tamanrasset, a doctor reported 355 cases over four years (2021) — single source, not compared.

WHAT IS CONTESTED

• The health impact itself. There is no peer-reviewed epidemiological study on cancers in the test zones. We documented twelve searches on international scientific databases: no result. And a leading Algerian oncologist, head of radiotherapy at the Pierre and Marie Curie Center in Algiers, says so publicly: “it is difficult to prove scientifically the link between recorded cancers and induced radioactivity”.

• Congenital malformations: asserted since 2011, never recorded in a census.

• The aquifers: measured in 1999 below detection limits, never re-measured since.

The documentary void is the subject here. It is neither France nor the associations that have produced a study: no one has done it. And that is precisely what Algeria is asking for.

⚖️ The compensation scheme: 73 requests, one granted

Here are the most telling figures of the entire dossier. They are official, they come from the Committee for the Compensation of Victims of Nuclear Tests itself, and they need no comment.

Exact formulation, which we stand by word for word: since 2010 and until the end of 2025, the Committee for the Compensation of Victims of Nuclear Tests registered 73 requests from people residing in Algeria. It accepted one. A second was obtained by court decision.

A mandatory precision, without which the figure would be misleading: the Committee counts by country of residence, and not by nationality. An Algerian residing in France is therefore counted as “metropolitan”. That is why the official figures and those of the Algerian associations diverge — and that is why we give the counting rule at the same time as the figure.

To situate the scale: between 2010 and 2025, the scheme paid €102,408,063 to 1,492 victims, all countries and all sites combined, an average of €69,490. The Committee does not publish the breakdown of these amounts by country: any assertion of the kind “France paid such a sum to Algerian victims” is therefore unverifiable, and we will not make it.

Why so few acceptances? One must know the history of the text, because it is edifying.

• The Act of January 5, 2010, known as the Morin Act (loi Morin), created the scheme. But it required that the risk attributable to exposure be “negligible” — that is, a probability of 1%. Result: between 2010 and March 2015, out of 862 decisions, only 17 approvals, or 2%.

• The “negligible risk” criterion was abolished by the Act of March 20, 2017. Result: between 2015 and 2024, out of 2,315 decisions, 959 approvals, or 41%.

• A threshold of 1 millisievert was then introduced — first by a resolution of the Committee itself, on May 14, 2018, described by the National Assembly as taken “by going beyond its own prerogatives and the Morin Act itself”. Five of the eight members of the Committee's board resigned on that occasion. The law intervened only afterwards, in December 2018.

One thing French law does NOT do, and which must be understood: the compensation scheme acknowledges no liability. The Council of State wrote it very clearly in an opinion of October 17, 2016: this regime does not have as its object to recognize that the State “would have the status of responsible party or of liable third party for the damage”. It is a regime of national solidarity, not of liability.

On the other hand — and this is a point that the persons concerned often do not know — it is NOT for the applicant to prove the causal link. The Council of State, in an opinion of June 28, 2017, specifies that the presumption can be rebutted only if the administration establishes that the illness “results exclusively from a cause extraneous to” the exposure. The burden of proof lies on the administration, not on the victim.

Two opposing court rulings, and they concern Algerians:

A victory. The Paris Administrative Court of Appeal, on December 27, 2017, ruled in favor of the widow of an Algerian born in 1926, who died of lung cancer in 2001, who had been “recruited locally by the directorate of special works of the engineering corps of Reggane” and had worked as a plumber at the Saharan center for military experiments between June 1959 and August 1961. The judges' reasoning deserves to be quoted: the tests “locally highly contaminated the sand”.

A defeat. The Council of State, on February 3, 2021, rejected the request of a man assigned to In Amguel, in Algeria, from June 1, 1962 to April 30, 1966: his dosimetric monitoring was judged sufficient, all results being below 1 millisievert.

🇺🇳 What France acknowledged before the UN — and what it did not say

This is the strongest element of the entire dossier. It is official, it is verifiable, and it is not polemical.

In 2023, during the Universal Periodic Review of France before the United Nations Human Rights Council, several recommendations were submitted on the victims of French nuclear tests. Notably:

• recommendation No. 45.222, formulated by Algeria: “Guarantee access to justice and appropriate reparation for all victims of French nuclear tests”;

• recommendation No. 45.220, formulated by the Marshall Islands, expressly targeting victims “including in North Africa”;

• recommendation No. 45.221, formulated by Panama.

France's response, in the official document, is this: “France accepts this recommendation.”

And here is the limit, which must be published at the same time, because it is decisive: the substantive justification presented by France speaks only of French Polynesia. In the entire French response annex, the word “Algeria” appears only once, and off the subject.

In other words: France formally accepted, before the United Nations, a recommendation covering all victims of French nuclear tests — including those of North Africa — while justifying its action only for Polynesia.

Three other institutional elements, which we give for completeness:

• In September 2024, UN special procedures addressed one communication to France and one to Algeria, mentioning 12,000 alleged victims. Formulation to be quoted as is: the rapporteurs write “without wishing to prejudge the accuracy of these allegations”. This is not a finding: it is a transmission of allegations. No response from either state is recorded to date.

• In 2015, the UN Human Rights Committee had declared itself “concerned at the 98.3% rejection rate as of March 1, 2015 of the files submitted to the compensation Committee”. The Committee does not name Algeria: it speaks of the French scheme as a whole.

• And, contrary to what is often read: in 1959, even before the first test, the UN General Assembly had adopted a resolution that “expresses its grave concern at the intention of the French Government to carry out nuclear tests; calls upon France to refrain from such tests”. These resolutions were not binding — and that is exactly what must be written: the international community had asked France to renounce them, without having the means to compel it to do so.

🚫 Four things we will not write — and why

We prefer to forbid ourselves these four assertions rather than see them demolished by others. Each is false, and each circulates.

First: “France violated the Moscow Treaty”. False, for two reasons. The treaty banning tests in the atmosphere, in space and under water dates from August 5, 1963 — after the Reggane tests. And above all: France was not a party to it. We checked the United Nations depositary registry, which lists 395 acts for this treaty: the word “France” does not appear in it a single time. The registry does, however, carry: “Algeria — Signature — August 14, 1963” — Algeria signed it and has never ratified it.

Second: “France acknowledged its responsibility”. False, and it is the French Council of State itself that says so (opinion of October 17, 2016, cited above).

Third: “150,000 Algerian victims”. This figure is real but diverted. In all the official documents where it appears — the impact study of the 2010 Act, an AFP dispatch from 2012, the National Assembly report of 2026 — it designates the potentially exposed population, all sites combined, Sahara AND Polynesia, military and civilian included. It is never a number of Algerian victims. We will not use it in that sense.

Fourth: “no UN body has ever taken a position”. Also false, as the previous section shows: the 1959 resolution, the observation of the Human Rights Committee in 2015, the recommendation accepted by France in 2023, the communications of 2024. The truth is more interesting than the exaggeration: the UN has taken a position several times — and Algeria was never named there by France.

🗄️ The archives: a new right that few people know about

Algeria demands three things, and they have been constant for fifteen years: recognition, decontamination, and the maps — that is, the exact location of the test sites and of the buried waste.

The Algerian requests, in chronological order:

• June 9, 2009 — Mourad Medelci, Minister of Foreign Affairs: complete mapping and clean-up, “and not to reduce the problem to a simple question of compensation”. This is the oldest documented request for mapping, and it predates the French Act of 2010.

• February 11, 2011 — Farouk Ksentini, president of the national consultative commission: he speaks of “crime against humanity” and asks that France “offer apologies” and compensate the victims.

• December 30, 2022 — President Tebboune: that France “clean up the sites” of these tests, “around Reggane and Tamanrasset, where the pollution is enormous”.

• December 29, 2024 — President Tebboune, before both chambers of Parliament: “come and clean up the contaminated zones. Do not give us money. I do not need your money.”

• February 2025 — the journal of the Algerian army sets out the three consolidated demands: recognition, decontamination, and the handover of the archives AND of the location of the waste burial sites.

• Late April 2026 — before the review conference of the Treaty on the Non-Proliferation of Nuclear Weapons, in New York, Ambassador Amar Bendjama publicly puts three questions that have remained unanswered, including this one: “Why has France still not provided Algeria with the maps indicating the exact location of the nuclear test sites and making it possible to locate the buried waste?”

What this means legally, and it is important: what Algeria is asking for is not money. In the vocabulary of international law, its demand falls under restitution — decontamination, the maps — and under satisfaction — recognition. Not under compensation. That is why the €102 million paid by the French scheme does not answer the Algerian demand: it compensates individuals, it decontaminates nothing and it acknowledges nothing.

And here is a lever that almost no one knows about. By two decisions of July 2, 2021, the French Council of State held that “archives that have been subject to a classification […] are communicable as of right upon expiry of the 50- or 100-year periods […] even where they have not been declassified”. In other words: since 2021, the “secret-défense” defense classification no longer constitutes, by itself, an obstacle to the release of archives more than fifty years old. That period has expired for all of the Algerian tests, which ended in 1966.

One exception remains, and it targets precisely the Algerian request: documents that could enable “the design, manufacture, use or location of weapons of mass destruction” remain non-communicable without any time limit. Location is expressly protected. There, in one sentence of law, is why the maps do not come.

Where are the documents? They exist, and they are catalogued. The Defense Historical Service holds two series specifically devoted to the Sahara:

• “NUCLEAR EXPERIMENTS IN THE SAHARA” — reference numbers GR/13/R/17 to GR/13/R/33, period 1957-1966;

• “NUCLEAR EXPERIMENTS from 1960 to 1966” — reference numbers GR/13/R/34 to GR/13/R/122, period 1957-1966.

We publish them because an Algerian reader has the right to know that these reference numbers exist. That obviously does not guarantee that he will obtain release.

Two findings, finally, that say a great deal:

• The archive-opening figures cited everywhere — 840 boxes analyzed, 398 freely communicable, 98,000 documents — concern POLYNESIA. There is no numerical equivalent for Algeria. And, as the director of the National Archives stated before a parliamentary committee on January 28, 2025: “the CEA has never transferred archives to the National Archives” — the French Atomic Energy Commission keeps everything itself.

• The commission for the opening of archives set up on October 5, 2021 concerns, in its very title, exclusively French Polynesia. No equivalent exists for Algeria.

🤐 Apologies: the answer is clear, and it is negative

We searched. Here is what we found.

Official French apologies to Algeria for the nuclear tests: none. Official regrets toward Algeria: none.

What exists is the speech delivered by President Macron in Papeete, on July 28, 2021. We counted the words in the official Élysée transcript, with the text de-duplicated — because the page contains the speech twice:

And the commission of inquiry of the National Assembly of June 2025 formulated forty-five recommendations. None is specifically devoted to Algeria — neither to Reggane, nor to In Ekker, nor to the mapping of the Algerian sites, nor to their decontamination, nor to their archives. The recommendation aimed at abolishing the millisievert threshold is explicitly limited “in French Polynesia”.

French parliamentarians themselves deplore it. In the Senate, on May 28, 2026, a senator declared: “We therefore simply ask that what was made possible for Polynesia be made possible for Algeria.” And the same day, two provisions of the text under discussion were withdrawn or rejected: the one that centralized the archives, and the one that provided for a report to Parliament on “a mapping of the known archive holdings” relating to Algeria. The only properly Algerian provision of the bill was rejected by the Senate.

⚠️ A contradiction we cannot settle — and we say so

Two official French sources contradict each other on the volume of archives processed, and we cannot reconcile them:

• the Defense Historical Service speaks of 840 boxes analyzed, of which 398 freely communicable — that is, about 98,000 documents — and 194 non-communicable documents;

• the Minister of the Armed Forces of the time, in a hearing, spoke of “173,000 documents” processed from “900 boxes”.

The two sets are not reconciled in the parliamentary report. We give both, without merging them. And let us recall it: neither of them concerns Algeria specifically.

🎯 What this dossier demonstrates — and what we ask for

We did not write this article to accuse. We wrote it to establish.

What the documents establish:

1. Seventeen tests, eleven of them after independence, the last on February 16, 1966 — it is the French National Assembly that says so.

2. The reference report on the sites, the IAEA's, contains no health data, describes itself as preliminary and acknowledges that its “limited and minimal” sampling restricts the validity of its conclusions.

3. No peer-reviewed epidemiological study exists on cancers in the test zones — and a leading Algerian oncologist says so publicly.

4. Out of 73 requests filed since 2010 by people residing in Algeria, one was accepted.

5. France formally accepted before the UN an Algerian recommendation on the victims of nuclear tests — while justifying its action only for Polynesia.

6. The Algerian archives exist and are catalogued, the release period has expired — but the “location” of the sites remains protected without any time limit. There, in law, is the answer to the question Algeria has been asking since 2009.

What we ask for, and we ask for it without insult:

That the maps be handed over. This is not a request for reparation: it is a request of public-health common sense. A country cannot protect its population from a risk whose location it is refused.

That an Algerian inventory be carried out. There is no official census of the Algerian workers who worked on these sites, and no official Algerian figure for the number of victims. The first step is not compensation, it is the census. It is moreover the only thing Algeria can do alone — and nothing prevents it.

That an epidemiological study be funded. We write this dossier from Algiers, and we observe this: as long as no one, in France or in Algeria, has produced a peer-reviewed study on these zones, the debate will remain an exchange of assertions. The void benefits whoever wants to do nothing.

And one thing that we say to our own camp: the exaggerated figures — the 150,000 victims, the 30,000 dead, the fallout over 3,000 kilometers — cost us more than they serve us. They allow our opponents to disqualify the entire dossier by correcting one line. The truth of Reggane is heavy enough that it needs no help.

🏛️ To decision-makers: the six decisions this dossier makes possible

This dossier is the heaviest we have published, and it is also the one where Algeria has the most levers in its own hands. Our correspondent wrote it from Algiers, after reading the French parliamentary reports line by line.

We are not asking to be listened to. We are asking for decisions. And it must be said plainly: four of the six decisions below depend on no foreign authorization.

One thing that we say to our own camp, and we stand by it. We set aside from this dossier the figures that favored us: the “150,000 Algerian victims”, the “30,000 dead”, the fallout “over 3,000 kilometers”. They are false, and they cost us more than they serve us — they allow our opponents to disqualify the entire dossier by correcting one line. The truth of Reggane is heavy enough that it needs no help.

The timeline, now. The inventory and the publication of the reports already transmitted can begin immediately. The epidemiological study requires a budget and three to five years. But a country that has not counted its own exposed workers, seventeen years after asking for maps, cannot reproach others for giving it nothing.

📚 All the sources — so that everyone can check

We ask no one to believe us.

Institutional texts and reports — International Atomic Energy Agency, November 1999 mission, report published in 2005 (full text consulted via the archive of the official link, today out of service); UNSCEAR, 2000 report, annex C, and 2024 report, annex B; Parliamentary Office for the Evaluation of Scientific and Technological Choices (OPECST), Bataille-Revol report, National Assembly No. 3571 and Senate No. 207, 2001-2002; Senate, report No. 18, 2009-2010 (impact study of the Morin Act); National Assembly, report No. 1558 of June 10, 2025 (commission of inquiry) and report No. 2364 of January 21, 2026; Senate, sitting of May 28, 2026; Committee for the Compensation of Victims of Nuclear Tests, activity reports 2024 and 2025; commission monitoring the consequences of the nuclear tests, report of July 3, 2007.

Law — Act No. 2010-2 of January 5, 2010; Decree No. 2014-1049 of September 15, 2014; Council of State, opinion No. 400375 of October 17, 2016 and opinion No. 409777 of June 28, 2017; Council of State, July 2, 2021, Nos. 444865 and 448763; Paris Administrative Court of Appeal, December 27, 2017, No. 17PA00415; Council of State, February 3, 2021, No. 439640; Council of State, February 24, 2022, No. 453671.

International statements and documents — United Nations General Assembly, resolutions A/RES/1379 (XIV) and A/RES/1402 (XIV) of November 20, 1959; Universal Periodic Review of France, 2023, report of the Working Group A/HRC/54/5, recommendations 45.220, 45.221 and 45.222; communications of the special procedures AL FRA 6/2024 and JAL DZA 2/2024 of September 13, 2024; Human Rights Committee, concluding observations CCPR/C/FRA/CO/5, 2015, § 21; Algerian statements at the IAEA General Conference of September 2025 and at the review conference of the Treaty on the Non-Proliferation of Nuclear Weapons, late April 2026; Élysée, Papeete speech of July 28, 2021.

Archives — Defense Historical Service, finding aids, series GR/13/R/17 to GR/13/R/33 and GR/13/R/34 to GR/13/R/122; Consultative Commission on the National Defense Secret, opinion No. 2013-07 of March 21, 2013; hearings of January 28, 2025.

Final note on method. We set aside from this article the following figures, after verification: non-institutional yield values; the “1,739 Algerian files” (the source link is dead, the official figure is 73); the “57 tests” (17 nuclear tests plus 40 complementary experiments, which must not be added together); the theoretical mortality coefficients that some present as estimates for Algeria, whereas they are calculated for one million people exposed to 10 millisieverts; and the assertion that the IAEA had “recorded high levels of persistent radioactivity” — the report says the opposite, except on three precise points, which we have mentioned.