Global nuclear weapon explosions
Several times already, I was surprised how little people seem to know about the global history and reality of nuclear explosions: many people know about the bombings of Hiroshima and Nagasaki, some (especially my dear cineasts) might also know about the first nuclear weapon test, named Trinity - but actually, since 1945, there were more than 2400 nuclear explosions [1] globally.
For some, those explosions signify important milestones for our global order relying on "nuclear peace", in which the great powers refrain from aggression, as this would simply be too dangerous with the existence of such enormous weapons. For others, these explosions mark the beginning of an age in which an order based on cooperation and peace got even more out of reach, as any interaction on the world stage has the threat of mutual destruction looming in the background. No matter the interpretation of that status quo - the stories of those affected by the nuclear explosions show us what terrible harm has been done, how it still effects the environment [2] and people, and how it will continue for quite some time.
There have been no nuclear explosions since 2017 [3]. We have a treaty banning all nuclear explosions, the Comprehensive Test Ban Treaty (CTBT), with an absolutely impressive global monitoring system that is running continuously to ensure that any nuclear test anywhere would be detected. Interestingly enough, the CTBT is not yet in force, however, the effect of the norm against nuclear weapons explosions seems to be.
In the last years, however, we have seen worrying developments: In 2020, Trump administration officials discussed whether the US should resume nuclear testing, and in Russia de-ratified the CTBT. In today's times, where norms seem to shift so fast, it is especially important to at least keep up the nuclear taboo. Looking at the data on nuclear weapons explosions might remind us why we cannot want them back.
The Johnston Archive – providing accessibility
One huge data base of global nuclear explosions is the Johnston Archive. The data base includes many parameters for each nuclear explosion, and as a data scientist by training, I was always tempted to spend some time diving into it; so far, I only had some oversight idea on nuclear weapons explosions.
Using that data base directly, however, is not too straight-forward, simply because of its raw format. In some cases, column positions vary within the table, or different columns overlap; those features complicate reading the data in an automated way. I decided to write some code to extract the data into formats that can be read in more easily (e.g. a html table and csv). I also included the data on the latest DPRK tests (for which only the first two are in the Johnston archive) from [4], and I added the geolocation information about the explosion sites (e.g., state, region) using the OpenStreetMap Nominatim API.
It would be great if the data in these formats would help those interested to engage with the topic and remove one barrier in learning about the past nuclear weapons explosions.
Global nuclear explosion visualisations
In the following, I share some interactive figures [5] which are the result of my engagement with the data base, purely following my own curiosity. They are not supposed to provide a thorough analysis of the historic global nuclear explosions but rather provide an invitation to explore the data - and maybe trigger some new questions, some new ideas or starting points to further engage with the topic.
Overview
The following pie charts are meant to provide an overview about the 2431 nuclear explosions from the data base. They answer the following questions: (1) Who ignited the explosions? (2) Where, i.e., in what region of the world [6]? (3) How, or in what medium/environment [7]? (4) Why, or what purpose did the explosion serve? (5) How big, or what explosive yields were detonated? The hover labels each indicate the absolute numbers within that category.
Who. Among the nine states, you might have noticed the category "Unknown*" with one nuclear explosion. This refers to the Vela incident, which is often attributed to a joint explosion by Israel and South Africa - the asterisk indicates that this is still affected by uncertainty.
Why. While the majority of explosions have been detonated for weapons purposes, some (which surprised me to see!) were meant for industrial applications, for which the data base even contains further sub-categories: cavity excavation (CV), earth moving (EM), extinguishing of oil/gas well fire (FE), oil stimulation (OS), and seismic sounding (SS). I did not include the sub-catagories into the pie chart, but if you're interested, feel free to explore them in the data base where you'll find them via the abbreviations I gave in brackets. To further improve readability, I collated the purposes joint verification (JV), military exercise (ME), combat use (C), vela uniform (VU), and storage/transportation experiment (ST) into "other". (Brackets show the abbreviations used in the data base).
How big. Choosing suitable numerical categories for the explosive yields, which are usually given in equivalents of 1000 tons of TNT ("kT"), is not an easy task - it should convey meaning, but how to do that if the explosive yields are simply too large to imagine. I decided roughly based on some benchmarks: a typical non-nuclear explosive has a yield of below 0.1 kT, the 2020 Beirut explosion was about 1 kT, the bombs exploded in Japan were of the order of 20 kT, the currently largest bomb in the US arsenal has about 1 MT (MT = 1000 kT), and the largest bomb ever tested was about 60 MT.
Where? Explosion locations
The following map shows where the nuclear explosions have occured, and the colors indicate what state ignited them. As there are some sites with more than one explosion, I added a yellow marker that indicates the places with multiple explosions. You can zoom in and hover on the points to get more information, i.e. the number of explosions at the site, the names of the explosions (to simplify finding the respective entry in the data base, if needed), the yield or yield ranges, the type (or explosion environment), the method (how the explosive was placed, e.g. "air drop"), the purpose, and the year (or year range).
Where? Aggregate numbers and yields per regions/ocean
This world map visualises the explosion numbers or yields summed up or aggregated for the different world regions or oceans (same regions/oceans as used above). The pie charts are roughly centered at the region or ocean it refers to (also visible in the hover label) and their sizes indicate the respective magnitude (aggregate number of yield), and the slices show the owners of the explosion. Using the buttons above the plot, you can switch between aggregate explosion numbers, aggregate explosive yield, and aggregate explosive yield when only considering atmospheric explosions (i.e., not taking into account the underground explosions).
While the US have performed most of their nuclear explosions in North America, the largest of their detonations (in terms of yield) were carried out in Micronesia; even more so, when looking only at the atmospheric explosions.
How high? - Height of burst
This plot shows the height of the nuclear explosion, again in color of the owner of the explosion. This plot was mainly inspired by a recent case of misinformation in which an AI system reacted to speculations about a nuclear explosion in Pakistan (definitely worth a read!).
It is clearly visible that after 1963, the explosions went mostly below the ground. This trend can be attributed to the Partial Test Ban Treaty, which went into effect in 1963 and banned all explosions besides those underground, and the ongoing debate and growing concern about the global radiological effects of atmospheric nuclear explosions, e.g., following the Baby Tooth Survey.
When? - Temporal changes and continuities with nuclear explosions
This histogram shows the number of nuclear weapon explosions for the different years. The buttons can be used to change the coloring, which indicates the distribution of explosion numbers across the following parameters: (1) states who ignited the explosions, (2) region of explosion location, (3) type of explosion, (4) purpose, (5) yield, and (6) method to place the bomb for explosion.
Note that you can click on the legend item to hide/show different categories, which is a nice tool to explore the data in this figure.
This figure is quite dense in information, and I guess that there is a lot which one could look at. For now, I'll just leave one observation, which somehow struck me: those unbelievably large weapons, with explosive yields larger than 10 MTs have existed since the 1960s - those weapons are technology from when the first video game was invented and the audio cassette was introduced… What old technology!
After 1976, however, those above-MT weapons seem to have vanished (at least in the available data). Today, most nuclear weapons states also abandoned those multiple-MT weapons from their arsenals. This means: there seems to be an upper limit on what destructive power is deemed to be useful. How is that assessment formed? It seems that the idea of which weapon is useful and which isn't is not set in stone - maybe, it is also on us.
Fin
The nuclear explosion data base contains even more information than what I touched upon this blog post. For example, there is also (some) data on the venting associated to the nuclear explosions, or on how large the crater associated to the explosion was, more information on the yield estimates, the exact timings… I hope this sneak peek has already sparked some interest; I would definitely be interested to see some more work on this. If you have any comments, questions or further ideas - let me know!