How USA TODAY mapped the potential consequences of a strike on US missile silos
The Nuclear Sponge is a five-part explanatory project by USA TODAY that aims to inform readers about the strategic debate and costs of modernizing the land leg of America’s nuclear triad.
The United States has 450 nuclear missile silos spanning six states in the Great Plains and Mountain West that hold 400 Minuteman III intercontinental ballistic missiles, or ICBMs. The Minuteman III can deliver a nuclear weapon virtually anywhere in the world within minutes.
But the Minuteman and its silos are aging. The infrastructure was largely built during the 1960s, and the Air Force used the Minuteman III during the early 1970s.
The Pentagon is working to replace the vintage ICBMs with a next-generation missile, dubbed Sentinel. However, developing the replacement nuclear missile and constructing new silos will cost taxpayers an estimated $140 billion, following major budget overruns.
Both supporters and opponents of the Sentinel program acknowledge that one of the reasons the U.S. has remote, land-based nuclear missiles is that an enemy trying to wipe out America’s nuclear arsenal would have to strike them with their own nuclear weapons, thus reducing the number of warheads falling on more populated areas of the country. This is called the “nuclear sponge” theory.
To inspire discussion of U.S. nuclear strategy and the Sentinel program, USA TODAY mapped the radioactive fallout that would follow an enemy first strike against America’s nuclear sponge. Here’s how we did it – the fallout model we used, the decisions and assumptions we made, other similar projects we consulted and the team that made it possible.
Why did our main story include just the nuclear sponge and not other targets?
That’s because we didn’t want our readers to become overwhelmed with a sense of fatalism when looking at thousands of nuclear detonations.
Our nuclear reality is the sum of many discrete policy decisions made by policymakers across the world’s nuclear states. One such policy decision is modernizing the land leg of America’s nuclear triad. The nuclear sponge can be understood and considered as a distinct piece of a larger puzzle. We believe in helping our audience understand the costs, stakes and realities of such choices.
The fallout model
For this simulation, USA TODAY used the WSEG-10 fallout model, which was developed in 1959 by George Pugh and Robert Galiano of the Institute for Defense Analyses think tank, under contract with the Pentagon’s Weapons Systems Evaluation Group. The original code was in the FORTRAN language. Air Force officer Dan Hanifen published an updated version as part of a master’s thesis at the Air Force Institute of Technology in 1980, which researcher Edward Geist (now at the RAND Corporation) ported to Python in 2016.
USA TODAY partnered with Alex Wellerstein, a nuclear weapons historian and professor at the Stevens Institute of Technology and visiting researcher at the Nuclear Knowledges group at Sciences Po, Paris, who ported WSEG-10 to Javascript for this project, largely based on Geist’s code, but with some alterations made with reference to Hanifen’s. (Wellerstein is also known as the creator of NUKEMAP, a free webapp that allows users to simulate the blast effects of a nuclear detonation.)
To make the visualizations and allow for experimentation and exporting of them, Wellerstein built “The Fallout Machine,” which merged WSEG-10 with custom target lists, integrated weather data and features, allowing the creation and export of the geographically-aligned fallout plume images that are used in this article.
The WSEG-10 model treats fallout deposition as a “smear” of radioactive material deposited downwind of the burst point. It was used extensively in the 1960s and 1970s by the United States government to model possible radioactive damage from a nuclear war. It is not the most sophisticated fallout model possible, but it is quick to calculate the effect of hundreds of overlapping plumes and does not require as much meteorological data as more sophisticated models do.
The model is regarded as a fairly “conservative,” according to Wellerstein, in that it guides planning for worst-case scenarios rather than erring on the side of underestimating the fallout. For any individual detonation, the WSEG-10 model is typically simpler than many other models; for large numbers of nearby detonations, such as an attack against the U.S. missile fields, the model’s results look very similar to those of other models, because the overlapping clouds tend to average out.
Wellerstein assembled the Fallout Machine’s weather conditions from average daily wind speed and direction data released by the National Center for Environmental Information, part of the National Oceanic and Atmospheric Administration (NOAA).
The NOAA data lacks a specific component required by the WSEG-10 model, horizontal wind-shear, which determines how wide the plume ends up being. For these examples, the Fallout Machine used Perlin noise to give smooth but random values within the range of likely values for this wind component. For the sake of simplicity, we assumed that all detonations occurred simultaneously.
Nuclear fallout is caused by radioactive material inside the mushroom cloud of a nuclear detonation falling from the sky as the cloud drifts with the wind in the minutes, hours and days after the explosion. Most of the radioactivity inside the cloud is a result of how much nuclear fission takes place, and thermonuclear warheads are usually assumed to have at least half of their explosive yield from nuclear fission (with the rest from nuclear fusion).
Nuclear detonations at or near the surface produce more intense fallout immediately downwind of them because their radioactive material becomes attached to heavier particles of dirt and debris inside the nuclear fireball. This is largely what is responsible for the “local fallout,” which is what most people picture when they think about fallout. It can cause acute radiation sickness and poses long-term contamination hazards. Some fallout particles remain aloft at high altitudes for a very long period, raising the overall global background radiation by a small but detectable amount.
Real-world fallout patterns are difficult to predict because they are largely dependent on the weather conditions, which vary at different altitudes of the mushroom cloud, and can change dynamically over time. Wind conditions also vary at different times of the year, and fallout can also come down “prematurely” because of precipitation (“rain-out”), including precipitation caused by the smoke from nuclear-weapon-induced burning.
All of this contributes to the difficulty of predicting the actual outcomes of nuclear detonations, and all the data scientists have on fallout patterns comes from the years of atmospheric nuclear testing. These tests were conducted as individual detonations (not hundreds or thousands of detonations), in only a few specific biomes (e.g., the deserts of Nevada, the coral atolls of the Pacific Ocean, the steppe of Kazakhstan, and the arctic archipelago of Novaya Zemlya), and generally under conditions thought to be highly predictable. The aftermath of a large nuclear attack contains many unknowns.
Building and simulating an enemy first strike
The WSEG-10 model requires a series of inputs to simulate a nuclear strike.
Location
First is the location of the detonation.
Minuteman silos are visible from space, and Russia also knows their locations thanks to data-sharing arrangements in arms control treaties. (The U.S. similarly knows where Russia’s missile silos are located.)
USA TODAY pulled the coordinates of 500 Minuteman silos from a 2013 United Nations report, “A New START Model for Transparency in Nuclear Disarmament: Individual Country Reports,” by researchers Tamara Patton, Pavel Podvig and Phillip Schell. The report lists the precise location of every known and suspected nuclear weapons site in the world; the U.S. sites are contained in Annex E.
Matt Korda, the associate director of the Nuclear Information Project at the Federation of American Scientists, assisted USA TODAY in identifying 50 silos near Malmstrom Air Force Base, Montana, that were deactivated in the late 2000s and later destroyed. That left USA TODAY with a list of 450 targets for our strike simulations.
Weapons parameters
Our WSEG-10 tool also requires a series of inputs about the simulated weapons used in the strike, most notably their yield, altitude of detonation, warheads per ground zero and their circular error probability (or accuracy).
USA TODAY consulted with experts (including Korda of FAS and Podvig of the UN) and reviewed academic and specialist literature to inform our decisions about weapons inputs. Our goal was to create a plausible enemy counterforce first strike against the U.S. missile fields. (Counterforce targeting, in nuclear weapons parlance, is a targeting strategy in which a country attacks the nuclear weapons of another country to prevent retaliation.)
The only country that currently has enough nuclear weapons to conceivably launch a counterforce strike against the United States is Russia, so USA TODAY based its weapons analysis on Moscow’s nuclear arsenal.
Minuteman nuclear missile silos are designed to resist (if not survive) a Russian nuclear attack, and experts believe that destroying a silo requires detonating at least one nuclear warhead. Many experts USA TODAY interviewed, including Kyle Balzer of the American Enterprise Institute and Bob Peters of the Heritage Foundation, argued that an adversary attempting to destroy the silos would aim two or more warheads at each launch facility.
Russia’s nuclear arsenal is currently undergoing a major modernization push, and Moscow’s war against Ukraine has led authorities to disclose even less information about the country’s nuclear weapons than in the past. Three of the best English-language sources for information about Moscow’s nuclear arsenal are: the annual Russia installment in the Nuclear Notebook by FAS, UN researcher Podvig’s Russian Nuclear Forces project, and the Center for Strategic and International Studies’ Missile Threat Project.
Ultimately, USA TODAY decided to simulate detonating one 800-kiloton warhead at ground level on top of each Minuteman III silo with a 220-meter circular error probability. This corresponds with the yield and accuracy of the SS-18 Satan ICBM, Russia’s venerable heavy nuclear missile, as described in archival documents translated by Podvig. The SS-27 Topol-M also carries a warhead believed to be of similar or larger size, though its accuracy is unknown.
Why did we only simulate one 800-kiloton warhead? USA TODAY assessed that if Russia were to attempt a counterforce strike against the United States, Moscow may reserve its (presumably) more accurate modern missiles for trickier targets such as command centers, communications nodes or similar facilities.
Weather
For the main story, USA TODAY simulated nuclear first strikes on three dates: April 1, 2024; April 27, 2024; and March 5, 2017.
We consulted with several climate and atmospheric scientists about the WSEG-10 model and weather in the missile regions on these three dates.
These scientists included: Josh Coupe, a postdoctoral researcher at the University of Colorado, Boulder who specializes in modeling nuclear and impact winters and their effects; Alan Robock, a distinguished professor of climate science at Rutgers University, who studies nuclear winter; Richard Turco, a distinguished professor emeritus and founding director of UCLA’s Institute of the Environment and Sustainability; and Brian Toon, a Professor emeritus at the University of Colorado and a Senior Research Scientist in the Laboratory for Atmospheric and Space Physics.
The weather patterns on April 27, 2024, were fairly typical for the nuclear sponge region. Coupe described the April cold front experienced that day as “pretty normal,” but noted that there was a strong high-pressure system over the eastern United States that day.
April 1, 2024, was even more typical, according to Coupe, who described the stationary front and cold weather across the region as a “very normal-looking day.”
USA TODAY intentionally selected March 5, 2017, as an outlier day with strong winds. Three massive wildfires, classified as “megafires” due to their large size, erupted across the southern Great Plains the following day due to the weather conditions. The largest of these fires, the Starbuck Fire, was, for a time, the largest wildfire in Oklahoma and Kansas history, according to the National Weather Service. Coupe emphasized that day had a “very strong cyclone and frontal system over the nuclear sponge.”
However, Coupe also noted that precipitation in the northern Great Plains could have made the radioactive material fall across those areas in unpredictable ways, a factor that WSEG-10 does not consider.
Possible shortcomings
It's important to mention some additional potential shortcomings to USA TODAY’s modeling approach.
Some researchers, notably Podvig, argue that Russia is unlikely to strike the nuclear sponge even in the event of a counterforce first strike. They question why Russia would shoot missiles at silos that might be empty by the time the strike arrives.
That’s because the United States is widely known to have a so-called “launch on warning” alert posture for its missile fields – if a Russian launch were detected, the United States would likely launch its ICBMs at Russian targets before Moscow’s first strike hit the ground.
“I think it's quite unlikely that Russia would target ICBM silos,” Podvig said in an email.
However, even if the Russians don’t use this strategy, U.S. strategists see targeting enemy silos as a key part of damage limitation: destroying any weapons that either weren’t launched or failed to launch in the first blows of a nuclear war prevents their later launch. That’s part of why the nuclear sponge theory is influential among U.S. nuclear thinkers.
There’s also the question of how many 800 kt warheads are currently available in Russia’s arsenal.
The SS-18 Satan is aging rapidly, though it is younger than the Minuteman III. Federation of American Scientists researchers assessed in 2025 that 34 Satans remain on operational duty. The missile carries a multiple independently-targetable reentry vehicle (or MIRV), meaning that one Satan missile can strike multiple targets. The Satan missiles currently in service carry 10 800kt warheads, meaning that the 34 missiles can hit 340 targets.
Russia’s SS-27 M1 ICBM, known as the Topol-M, also carries a single 800kt warhead according to FAS and Podvig. Moscow has 78 such missiles in operational status, according to FAS. Adding the Satans and Topol-Ms, Russia has at least 418 800kt warheads available for launch, according to experts. That is fewer than would be needed to hit the 450 Minuteman silos.
USA TODAY recognizes that our 800kt per U.S. silo scenario thus does not perfectly align with Russia’s present-day arsenal. Korda of FAS suggested a conservative strike scenario would instead allocate two 100kt warheads – an explosive yield carried by Russia’s newest ICBMs and its submarine-launched ballistic missiles – per Minuteman silo.
We made this decision for a few reasons. One is the nature of the FAS estimate of Russia’s Satan missile force: its 2025 report noted that Moscow may have 40 Satan missiles on alert rather than 34, which would provide an additional 60 warheads for our scenario.
Another is that Russia is developing a new heavy ICBM – the SS-29 Sarmat – to replace the Satan. It’s not yet clear when the Sarmat will be combat-ready after a series of failed tests, but the project shows Moscow’s commitment to keeping high-yield warheads in its nuclear arsenal.
Moreover, the fallout model generates similar, if somewhat less severe, results for a strike conducted with two 100kt warheads per silo rather than one 800kt warhead. We’ve embedded images below comparing the 800kt strikes from the main graphics story with double 100kt warhead strikes.
Additionally, one of the most significant interactive research projects on fallout from a strike against the nuclear sponge also assumes strikes with 800kt warheads.
Other models and research
That project, Missiles on Our Land, was created by a group of researchers at Princeton University and Columbia University and published online and in Scientific American.
The lead researcher for Missiles on Our Land was Sébastien Philippe, a nuclear risks researcher who has since moved to the University of Wisconsin, Madison and received a MacArthur Fellowship – better known as a “genius grant.”
Philippe’s project, published in 2023, also projected attacks against the nuclear sponge. His team built its own sophisticated fallout model and simulated 365 attacks – one based on the weather conditions for every day of 2021.
One of its products was an interactive map, Under the Nuclear Cloud, which visualizes the average risk of radiation exposure across 365 simulations for most of North America. USA TODAY encourages readers interested in assessing the fallout risk in their own communities to consult Philippe’s work.
Other researchers, including some of the climate scientists mentioned above, also specialize in modeling the effects of nuclear fallout and smoke from nuclear detonations. Their methods are more academically rigorous than ours, and USA TODAY recommends interested readers consult the academic literature on the subject.
Project credits
USA TODAY’s Nuclear Sponge series was conceptualized and designed by defense reporter Davis Winkie.
Winkie reported and wrote the majority of the project’s five stories with significant contributions from visual journalist Ramon Padilla and senior graphics editor Shawn Sullivan. Winkie also curated the archival and wire photographs included in the stories.
Washington editor Gabrielle Banks edited the text for all five of the project’s stories. Melissa Galbraith and Robert Abitbol provided copy edits.
Alex Wellerstein developed the Fallout Machine app, which generated the project’s fallout maps. Winkie, Padilla and Sullivan provided Wellerstein with feedback throughout the development process.
Padilla and graphics reporter Karina Zaiets designed the webpage for Part 1 of the project, using tools such as HTML, Svelte, Javascript, Geojson, Mapbox and Mapbox GL to put the fallout maps onto the page. Sullivan and graphics managing editor Alberto Cuadra contributed both hands-on and developmental editing for the webpage, and Winkie and Banks also contributed developmental edits.
Winkie developed the weather and strike scenarios for Part One, with assistance from Sullivan and graphics reporter Carlie Procell.
Graphics reporter Stephen Beard created 3D graphics of the Minuteman III ICBM and its infrastructure with research assistance from Winkie.
USA TODAY’s audience team, in concert with Winkie, Banks, Padilla and Sullivan, designed the project’s rollout on the newspaper’s website and social media channels. Key contributors included visuals designer Alan Nguyen, audience editor Kevin Quinitchett and deputy audience editor Alex Connor, deputy audience editor Jay Cannon and managing audience editor Ashley Lewis.
If you wish to provide feedback on the project, you can reach lead reporter Davis Winkie at [email protected].
★★★
Davis Winkie's role covering nuclear threats and national security at USA TODAY is supported by a partnership withOutrider Foundation andJournalism Funding Partners. Funders do not provide editorial input.