Interactive Map of Nuclear Explosions, 1945–2017
Ash Tales
More than two thousand nuclear detonations are documented in this map. They begin with the Trinity test in New Mexico on 16 July 1945, include the wartime bombings of Hiroshima and Nagasaki, and continue through the six nuclear tests confirmed in North Korea between 2006 and 2017.
But “2,052 explosions” is not the same thing as “2,052 nuclear tests”. A single test operation could involve more than one device, historical records do not always agree, and reported yields are sometimes estimates or broad ranges. This interactive atlas is best read as a map of records in a particular public archive, not a perfect inventory of every nuclear event.
Explore the nuclear explosion map
Use the filters to isolate a responsible state or test environment, drag the year control to travel through the nuclear age, and switch from individual markers to the density view. In detonation mode, marker size is scaled logarithmically to the record’s upper reported yield; it does not represent blast radius or fallout.
ASH TALES // PUBLIC DATA EXPLORER
The nuclear age, mapped
Records by responsible state
Records by test environment
Records by year
What the map reveals
The first impression is not an even scatter across the globe, but a series of dense test landscapes: Nevada in the United States, Semipalatinsk in present-day Kazakhstan, Novaya Zemlya in the Russian Arctic, Lop Nur in China, the Algerian Sahara, and Pacific atolls including Bikini, Enewetak and Mururoa.
That geography also exposes an important distinction. The chart groups records by the state responsible, while each marker sits at the place of detonation. France’s 208 records, for example, include tests in Algeria and French Polynesia. The United Kingdom’s 45 include sites in Australia and the Pacific. A “country count” is therefore not a count of explosions inside that country’s modern borders.
The archive attributes 1,031 records to the United States and 714 to the Soviet Union. Together they account for roughly 85% of all records in this combined dataset. France follows with 208, then the United Kingdom with 45 and China with 43. India has three archive records, Pakistan two, and the six-event supplement covers North Korea.
Those totals should not be used as a simple league table of nuclear capability. They reflect different entry dates, testing strategies, definitions and surviving documentation. They count detonations in the source, not warheads built, stockpile size or destructive capacity.
The peak—and the move underground
The busiest year in the archive is 1962, with 177 records. The United Nations’ broader history describes 1962 as a year of extraordinary testing intensity, but even reputable sources differ by one or more events because they may count tests, devices or explosions differently. Here, 177 is specifically the number of rows in the FOA/SIPRI-derived dataset.
Public concern about radioactive fallout helped drive the 1963 Treaty Banning Nuclear Weapon Tests in the Atmosphere, in Outer Space and Under Water, commonly called the Partial Test Ban Treaty. It did not ban underground testing. The pattern in this archive changes dramatically: 347 of its 409 atmospheric records occurred by the end of 1963, while 1,385 of its 1,526 underground records occurred later.
That transition reduced the routine production of atmospheric fallout, but “underground” should not be read as “safe”. Underground explosions can vent radioactive material if containment fails, disrupt geology and leave contaminated test sites. They are less spectacular on camera; they are not consequence-free.
The test rate declined sharply after the Cold War. The Comprehensive Nuclear-Test-Ban Treaty was adopted in 1996 and established a powerful global norm against testing, but it has still not entered into force because the required Annex 2 states have not all ratified it. North Korea’s 3 September 2017 detonation remains the most recent confirmed nuclear test; the CTBTO lists six DPRK tests between 2006 and 2017.
Yield: useful, dramatic and uncertain
The source reports a lower and upper yield for most records, in kilotons of TNT equivalent. Where both are present and disagree, the information panel shows a range. Marker sizes use the upper figure on a logarithmic scale so that low-yield tests remain visible beside multi-megaton devices.
Two events dominate the upper end. The Soviet Union’s 1961 test usually known as Tsar Bomba is recorded at 50,000 kilotons, or 50 megatons. The United States’ 1954 Castle Bravo test is recorded at 15,000 kilotons. Bravo’s yield greatly exceeded its designers’ prediction and spread fallout across inhabited parts of the Marshall Islands.
Yield alone cannot tell you the area destroyed or the radiation hazard. Burst altitude, weather, terrain, weapon design and the meaning of the original estimate all matter. For that reason, these circles are comparative symbols—not a simulation. If you are looking for practical context rather than historical analysis, read our guide to what to do in a nuclear attack and the interactive guide to UK fallout shelters near you.
Where the data comes from
The core consists of 2,046 rows downloaded from Virginia Tech’s CORGIS nuclear explosions dataset. CORGIS republished a cleaned version of data derived from Nils-Olov Bergkvist and Ragnhild Ferm’s Nuclear Explosions 1945–1998, produced for the Swedish Defence Research Establishment and Stockholm International Peace Research Institute. The original report and catalogue record are available through the US Department of Energy’s OSTI archive.
Because that archive ends in 1998, this page adds six confirmed North Korean tests from the CTBTO chronology. They share an approximate Punggye-ri test-site coordinate. We deliberately leave their yields blank: public estimates vary, and combining them as if they were measured on the same basis as the older archive would create false precision.
The land outline is Natural Earth’s public-domain 1:110 million dataset. Both the event data and basemap are stored with this site rather than loaded from a third-party chart provider, so the article does not depend on a deleted dashboard or expiring embed.
Method and limitations
- Each point is one row in the combined dataset. Multiple rows can share a date and coordinate.
- Dates follow the archive’s recorded time, commonly UTC, so a date can differ from the local calendar date at the test site.
- Coordinates and yields may be approximate. Some Soviet records have especially wide yield ranges.
- “Responsible state” follows the historical source; “Soviet Union” is retained rather than reassigned to modern states.
- Raw detonation types are grouped into five readable environments: underground, atmospheric, surface, underwater and space.
- The density layer counts records in screen-sized grid cells. It is not radiation, fallout or present-day contamination.
- The map does not include nuclear accidents, reactor disasters, subcritical experiments or unconfirmed claims because they are not nuclear explosions in this source.
No public historical dataset this large is beyond dispute. The value is in making its definitions visible, keeping the source traceable and allowing readers to inspect the pattern rather than simply repeating a total.
Reading beyond the map
For the human history behind these points, see our reading lists on nuclear war and the atomic bomb. The map shows the scale and geography of the nuclear age; those accounts show what its abstractions meant to the people beneath them.