In 1983, atmospheric scientist Carl Sagan and four colleagues published a paper in Science arguing that even a relatively small nuclear war could trigger global climate cooling so severe that crops would fail worldwide. The paper became known as the "TTAPS" study after the authors' initials. Its central claim — that soot from burning cities could darken the planet for years — was controversial at the time, but recent modeling has confirmed and in some ways strengthened the original conclusions.
The TTAPS hypothesis
The TTAPS authors observed that nuclear weapons detonated over cities would ignite enormous fires. Modern cities contain large amounts of fuel — wood, plastics, asphalt, vehicle fuel — and a single megaton-class detonation could ignite a firestorm covering tens of square kilometers. The resulting plume of soot, lofted by the heat into the upper troposphere, would drift on winds and could be transported globally by stratospheric circulation.
The original 1983 paper estimated that a 5,000-megaton war (roughly the deployed yields of the time) would inject 200 million tons of soot into the upper atmosphere. The soot would absorb sunlight, dropping global average surface temperatures by 15-25 °C. It would also remain in the stratosphere for years, because at those altitudes there is no rain to wash it out. Crops would fail across most of the Northern Hemisphere.
Critique and refinement
The TTAPS paper drew immediate criticism. Defense-aligned analysts argued that the soot loading was overestimated, that smoke would not reach the stratosphere, that smoke would rain out faster than predicted, and that warming caused by reduced cloud cover might compensate. By the late 1980s, a more cautious "nuclear autumn" framing had partly displaced the original "nuclear winter" claim.
Subsequent modeling — particularly the work of Alan Robock and Owen Brian Toon and collaborators in the 2000s and 2010s — addressed these critiques using modern coupled climate models. The results were closer to TTAPS than to the more cautious revisions: soot from city firestorms does reach the stratosphere, does persist for years, and does cause significant cooling. The original hypothesis has held up remarkably well.
The mechanism, in more detail
The mechanism is straightforward but not intuitive. A nuclear weapon detonated over a city ignites multiple fires that merge into a firestorm — a self-sustaining fire so intense it generates its own weather. The firestorm produces a vertical convective plume that lofts smoke and soot to altitudes of 10-15 km, just at the boundary between troposphere and stratosphere.
In the stratosphere, soot absorbs sunlight and re-radiates it as heat. This warms the stratosphere and cools the surface — the opposite of greenhouse gases, which cool the stratosphere and warm the surface. The cooler surface produces less convection, weaker monsoons, and shorter growing seasons. The stratospheric soot is removed only by gradual gravitational settling (months to years) since stratospheric circulation does not include rain.
The total effect depends on how much soot reaches the stratosphere. Modern modeling suggests that 100 city firestorms could inject enough soot to drop average surface temperatures by 5 °C for several years, with much larger drops in mid-latitudes during growing season.
Even a regional war causes global cooling
The most striking result of recent modeling is that even a regional nuclear war — say, between India and Pakistan — could cause significant global climate disruption. Robock and Toon's 2007 paper modeled an India-Pakistan exchange of about 100 Hiroshima-yield weapons (15 kt each) over each side's major cities. The result: 5 million tons of soot lofted to the stratosphere, and global surface cooling of 1.25 °C lasting about a decade.
A 2014 follow-up using improved climate models found even stronger effects: ozone depletion of 30-50%, growing-season shortening by 10-40 days across the Northern Hemisphere, and global crop yield reductions of 10-40%. The conclusion is that a "regional" nuclear war is not really regional — its climate effects affect the entire planet.
Crop failure and famine
Research by Lili Xia, Alan Robock, and others in the late 2010s and early 2020s extended the modeling to global food production. The conclusions are sobering: even a regional India-Pakistan exchange could reduce global wheat, corn, and soybean production by 10-15%, sufficient to trigger food shortages, price spikes, and famine in regions that depend on food imports. A larger US-Russia exchange could reduce global production by 50% or more — and the resulting global famine would likely kill more people than the bombs themselves.
The 2022 paper by Xia et al. in Nature Food estimated that a US-Russia nuclear war scenario could lead to over 5 billion deaths from famine in the years following the exchange — far more than the few hundred million who would die from the immediate effects of the bombs.
Why the implications matter
Nuclear winter changes the strategic calculus of nuclear weapons in important ways. A "limited" nuclear exchange may not be limited in its consequences. A nuclear war between two countries on the other side of the world could cause harvest failures in countries that have nothing to do with the conflict. The traditional concept of nuclear sovereignty — that a nuclear-armed state's decisions are its own — breaks down when the climate effects are global.
The simulator does not model nuclear-winter effects directly. Its calculations are limited to the immediate physical effects of single detonations. But for context: the global consequences of even a regional nuclear war would, on current modeling, dwarf the local effects shown in any individual scenario page on this site. See the resources page for nuclear-disarmament organizations and the sources page for primary references.