Nuclear Winter: How Much Do We Really Know?

A well-publicized study predicts that even a limited nuclear war would plunge much of the earth into darkness and freezing, threatening the very survival of mankind. Many are taking the prediction as fact, but uncertainties abound.


Modern intellectual history is replete with examples of the use of incomplete, inconclusive science for political ends. Rachel Carson's Silent Spring, published in 1962, scared a generation into believing that herbicides and pesticides were causing wholesale poisoning of the environment and that stringent government regulation was the only solution. Her disciple Barry Commoner translated environmental concern into ideological terms, asserting that the underlying source of pollution is capitalism itself, as if socialist farmers don't need fertilizer and weed killer to increase crop yields, or socialist factories don't emit smoke.

The doomsaying Stanford ecologist Paul Ehrlich and the Club of Rome, an informal group of thinkers concerned with global economic problems, published studies claiming that severe government intervention—perhaps even totalitarian methods of population control—were necessary to prevent "the population bomb" from destroying civilization. It is now nearly a generation later, and the apocalyptic predictions are nearly forgotten.

In a similar vein, Samuel Epstein in The Politics of Cancer (1978) summarized the misinformation propagated by a small group in the National Cancer Institute, claiming that industrial chemicals were generating an epidemic of cancer in the United States, and again (surprise) the solution was to be yet more government regulation. Only recently, Edith Efron exposed the falsity of this claim in her brilliant book The Apocalyptics: Cancer and the Big Lie (1984).

The latest and most pervasive episode of the political misuse of science is the prediction of global climatic catastrophe from nuclear war, even if a small portion of the nuclear weapons in the world are ever exploded. The apocalyptic idea is that nuclear explosions and the fires they would ignite would send dust, smoke, and soot high into the atmosphere, where they would block the sun's heat and light from reaching the earth's surface, thus plunging the world into darkness and subfreezing temperatures: crops would fail, creatures would die, humankind itself might perish. Propagators of this idea are using it to promote various political causes, including "the nuclear freeze," unilateral disarmament, and proposals for immediate, drastic bilateral weapons reductions to a level below which catastrophe is felt to be unlikely.

Of course, there must be a buzzword for the publicity campaign and a "prophet of doom" for TV appearances. The buzzword is nuclear winter, and the leading prophet is Cornell University astronomer Carl Sagan. Sagan is the star of a media campaign that opened with his article "The Nuclear Winter," published in the Sunday newspaper-supplement magazine Parade on October 30, 1983. The article appeared on the eve of a two-day conference in Washington, D.C.—the Conference on the Long-Term Worldwide Biological Consequences of Nuclear War, in which celebrity-scientists Sagan and Ehrlich were chief participants.

The well-orchestrated publicity campaign actually began in April 1983, when the conference organizers hired the public-relations agency Porter & Novelli Associates to publicize the issue and the impending conference. In an interview in Science Writers, a publication of the National Association of Science Writers, Porter & Novelli chairman Jack Porter unabashedly laid out the nuclear winterists' publicity strategy, reporting that the campaign ended up costing $100,000. Detailing the nuclear-winter PR tactics—including, for instance, efforts to arrange coverage by such major media as the Phil Donahue Show and Parade magazine—Porter recalled how the bombing of the Marine quarters and especially the US invasion of Grenada, though initially perceived as publicity "competition" (because both events occurred shortly before the conference was held), actually helped "heighten the dramatic perception" of the nuclear-winter idea: reporters returning from Grenada "quickly ran out of news," because of the restrictions on covering the invasion, Porter surmised, "and they had to come back and write about something else. Our story formed a kind of logical focus."

The first article in a scientific journal to spell out the gloomful hypothesis was "Nuclear Winter: Global Consequences of Multiple Nuclear Explosions," by Richard Turco, Brian Toon, Thomas Ackerman, James Pollack, and Sagan, published in the December 23, 1983, issue of Science. (This article is often called TTAPS, referring to the first letter of each author's name.) This sequence of events—a publicity campaign paid for and launched before the publication and circulation of a scientific study—is very unusual. In fact, most scientists agree that this type of arrangement is destructive of the goals of honest inquiry and more consistent with attempts at stock-market manipulation or disguised political purposes.

Following the December 1983 Science article, additional doomful publications appeared, including "The Climatic Effects of Nuclear War," again by the TTAPS authors (Scientific American, August 1984). Transcripts of the proceedings of the Conference on the Long-Term World-Wide Biological Consequences of Nuclear War came out in a book, The Cold and the Dark (1984), by Paul Ehrlich, Sagan, Donald Kennedy, and W.O. Roberts.

Also following the publication of TTAPS, other studies of one aspect or another of the nuclear-winter hypothesis have been done, and still others are planned. For example, the National Research Council—principal research arm of the renowned National Academy of Sciences—performed a study of the hypothesis, the results of which were issued in December 1984. Moreover, the issue immediately percolated into the mainstream media, with scores of newspaper reports, magazine articles, and TV-show segments bringing the gloomy news to the American public.

The beginning of the nuclear-winter idea is attributed to Dutch scientist Paul Crutzen, director of the Max Planck Institute for Chemistry in West Germany. He had been concerned about nuclear war depleting ozone in the stratosphere, the uppermost part of the atmosphere. In 1982, he and a colleague, American chemist John Birks, predicted that widespread nuclear explosions could ignite enough forest fires to produce a thick smoke layer that, for a period of months, would reduce the amount of sunlight reaching the earth's surface. Upon reading Birks and Crutzen's study, Richard Turco, a young atmosphere scientist with R&D Associates in Los Angeles, realized that the loss of heat from sunlight could cause a severe temperature drop. Turco is credited with coining the term nuclear winter. Turco joined with Sagan and three scientists from NASA's Ames Research Center in California—Toon, Ackerman, and Pollack—to devise the models, assumptions, and calculations for the study that bears the acronym TTAPS.

Soon after the TTAPS article was published in Science, Sagan extrapolated the policy implications of the nuclear-winter predictions in a paper in the influential quarterly Foreign Affairs. He advocated rapid decreases in worldwide nuclear armaments to a total of fewer than 1,000, a "threshold" below which Sagan believes that nuclear winter is very unlikely.

The nuclear-winter idea has predictably been taken up by various groups advocating unilateral nuclear disarmament, accommodation with the Soviet Union, and the nuclear-freeze movement. Civil defense, the favorite whipping boy of the antinuclear factions, comes in for the usual attack: Why bother with shelters when we'll all freeze or starve?

In reality, the implications of the nuclear-winter idea are the opposite of what Sagan and his followers claim. Nuclear winter is avoidable, even in case of major conflict using nuclear weapons. Because of the possibility of disastrous climatic effects from attacks on cities—which, when burning, produce more smoke and soot than other targets—US military planners will continue to shift their target priorities from population centers (cities) to militarily strategic sites, as Soviet planners are believed to have already done. This shift will result in greater need for civil defense, fallout shelters, food reserves, and medical preparations. A "freeze" on current weapons, even if it were mutually verifiable, would be the worst possible response to the possibility of a nuclear winter, because older-generation weapons now present in the Soviet arsenal are the riskiest for causing severe atmospheric effects.

Indeed, fear of a nuclear winter is more likely to encourage further efforts toward nuclear-weapon modernization than toward general nuclear disarmament. That "conventional" nuclear weapons might cause climatic catastrophe even reinforces arguments for the "neutron bomb," which produces mostly temporary radiation instead of blast and fire. Furthermore, the possibility of a nuclear winter may make more urgent the need for active defense measures such as High Frontier and the Strategic Defense Initiative ("Star Wars"), the proposals for space-based defensive systems against nuclear attack. Even if they are not 100 percent effective, these could drastically reduce population destruction and climatic effects.

These implications may seem paradoxical, but think again—paradox abounds in the Nuclear Age. That a Strategic Arms Limitation Talk (SALT I) could intensify the arms race by causing multiple-warhead missiles (MIRVs) to be added to arsenals is a paradox. That the tragic attacks on Hiroshima and Nagasaki, with the resulting fear of nuclear weapons, may have prevented further nuclear war for 40 years is a paradox. There are more to come.

In order to understand the TTAPS study's nuclear-winter predictions, you must know something about its assumptions and method of calculation. Turco and company started with a series of "nuclear exchange" scenarios, in which they assume that a certain number of nuclear warheads explode with a given total explosive yield, measured in megatons (millions of tons of TNT equivalent, abbreviated "MT").

In some scenarios, a higher percentage of the explosions occur in surface bursts—in which weapons explode on or near the ground—on military targets (for example, missile silos and submarine or bomber bases); this has been called a "counterforce" strike. Other scenarios, known as "countervalue," assume a higher proportion of air-burst explosions, over cities.

Surface bursts form craters, generating a lot of fine-grained dust from explosive breakup, vaporization, and recondensation of the surface materials. Air bursts, on the other hand, make little dust but can cause instantaneous widespread fires: thermal radiation from the weapon's fireball ignites combustible materials, such as wood houses, dead trees, brush, and building roofs and contents.

To provide an example of what they perceive as a likely nuclear exchange and its climatic effects, the TTAPS authors used a "baseline scenario" assuming a total explosive yield of 5,000 MT from 10,400 explosions of various sizes. A majority of the explosive yield would be used in surface-burst attacks on military targets with one-fifth directed in "countervalue" attacks on urban or industrial areas. In another scenario, for example, TTAPS assumed that 100 megatons of explosives are distributed in 1,000 explosions, all in countervalue air bursts over cities, with no counterforce surface bursts on military targets.

Using published data from nuclear tests, TTAPS then calculated the amount of dust generated by surface bursts and assumed a distribution of dust-particle sizes to determine the amount of dust that is spread high into the stratosphere—smaller particles are carried higher and stay up longer. They calculated the amount of smoke nuclear explosions cause by multiplying (1) the area presumed ignited by thermal radiation from air bursts by (2) the amount of flammable material present per unit area, times (3) a factor for how much of this material is burned, times (4) a factor for the proportion that is emitted as smoke. They then extrapolated from previous fire events how high the smoke might rise in the atmosphere. They assumed that 5 percent of the urban fires become firestorms with smoke plumes as high as 19 kilometers (about 60,000 feet). Smoke and soot are highly critical elements of the nuclear-winter hypothesis (and therefore, too, the amount of smoke-producing burning that a nuclear conflict would ignite), because the black, carbon content of smoke and soot is highly absorbent of sunlight. Ordinary dust, on the other hand, tends more to scatter light than to absorb it.

Next TTAPS calculated the distribution of dust and smoke over time and their spatial distribution (how much of what size particles, at what altitude, in what area). This is an extremely difficult problem, involving assumptions of how a war might be fought—including targeting patterns and the timing of attacks and counterattacks—meteorologic conditions, characteristics of dust and smoke clouds, rates of mixing of particles in the atmosphere, rates of removal of particles from the atmosphere by "washout" from natural processes (such as rain), etc.

The TTAPS authors side-stepped the enormous complexity of this problem by assuming that the spatial distribution would be completely uniform over the Northern Hemisphere and, with regard to time, by assuming that the uniform universal "cloud" appears instantly at the beginning of a nuclear conflict.

Determining the cloud's absorption of the sun's rays involves assumptions about the optical properties of the cloud's smoke and dust particles. It also requires a model to calculate how much light and heat is transmitted to the earth's surface. TTAPS chose a "one-dimensional" model of light and heat transmission, which assumes that altitude is the primary (and only) variable that affects the radiation of light and convection of heat. Programmed with this model, computers can predict how temperatures at land surfaces will change over time after detonation of nuclear weapons.

For their 5,000-MT, baseline scenario, TTAPS predicted that the Northern Hemisphere average temperature would drop to -23 degrees Centigrade (-9 degrees Fahrenheit) after three weeks and only rise above freezing (32 degrees F) after three months. The prediction from the 100-MT scenario, in which only cities are attacked, is only slightly less severe, with two months of subfreezing temperatures. This seems surprising by comparison to the 5,000-MT, baseline scenario—which has 50 times the explosive yield—until one realizes that Turco et al. assumed the number of air-burst explosions over cities to differ by only a factor of two (1,000 in the 100-MT scenario versus approximately 2,000 in the baseline version). Moreover, TTAPS unaccountably assumed a doubling of the combustible material in the 100-MT (versus 5,000-MT) scenario, and they also increased the smoke-emission assumption for this case. The crux is that the explanation for the short-term temperature drop is dominated by the smoke from the burning of cities, with a much smaller effect from the long-term circulation of fine dust in the stratosphere, as produced by the surface explosions inherent in attacks on military targets.

In critically analyzing the TTAPS study's method and assumptions, it should first be acknowledged that the study is a brilliant tour-de-force, opening up whole new areas of scientific speculation and contention. The TTAPS authors bridged previously uncrossed gaps between nuclear-weapons physics, pyrology, aerosol dynamics, optical physics, thermal transport phenomena, meteorology, computer modeling of climate, etc., as well as nuclear strategy. They have been "thinking about the unthinkable" in ways that had escaped the late futurologist Herman Kahn and nuclear physicist and renowned weapons expert Edward Teller.

The climate effect of nuclear war is an important, perhaps a vital, question for all of us who live with this Sword of Damocles overhead. Nevertheless, when apocalyptic claims are being made that affect the psychological security of millions of people, when policy recommendations are made that can affect the freedom and physical security of hundreds of millions, the ideas, assumptions, and calculations must come under heavy scrutiny. It is in this spirit that I will consider the weaknesses of the nuclear winter prediction.

First, the TTAPS team's assumption that all explosion products—dust, smoke, soot, and other particulate—are spread uniformly and simultaneously over the surface of the hemisphere is physically impossible and not even a good approximation of any realistic situation. The TTAPS authors assumed, by analogy with known atmospheric effects of volcanic eruptions and cooling effects of dust storms on Mars, that both dust and smoke would be spatially well-mixed.

But in reality, the rapidity and uniformity of mixing depends on the distribution of particle sizes. And one would expect "nuclear" dust and smoke to have a distribution different from what is known from previous non-nuclear explosions.

The TTAPS researchers used this model ostensibly because it fit into their "computer code" and is simple enough to make calculations convenient. But the value of their results suffers thereby, for this model is equivalent to a worst-case assumption.

Intuitively, we know that any "break in the clouds" will allow much more sunlight at the earth's surface. Indeed, it can be proven rigorously that any nonuniformity in the spatial distribution (that is, patchiness in the smoke cloud) will allow more solar energy transmitted. We also know that nuclear explosion patterns and smoke plumes will be intrinsically nonuniform; thus, surface temperature drops and the darkness of the nuclear cloud's overcast will be less than the TTAPS model predicted.

Furthermore, the TTAPS team's assumption of simultaneity—that all dust and smoke would be injected into the atmosphere at once—is also a worst case. Explosions and subsequent spreading of smoke from a massive nuclear exchange are more likely to occur spasmodically and unevenly over a period of days to weeks. Hence the sharp maximum temperature drop that the TTAPS researchers calculated actually would be "smeared" over time, resulting in smaller net maximum temperature changes than the researchers predicted.

Without explicit justification, the TTAPS authors assumed that a firestorm (a devastating type of urban conflagration with especially powerful updrafts) would be a factor in 5 percent of cities burned and that smoke and soot from these firestorms would rise into the stratosphere (up to 19 kilometers, or about 60,000 feet) and stay there a long time, causing great cooling effects. In fact, most experts believe that firestorms, as distinguished from mass fires, are unlikely from nuclear explosions, because the special circumstances needed to generate a firestorm—a highly concentrated fuel supply and weather conditions allowing the fire to consume all available oxygen—would not be present and because the powerful blast wave that follows the thermal flash would blow out many nascent fires. There were massive fires at Hiroshima and Nagasaki in the blown-down rubble of buildings, but no firestorms. Furthermore, smoke plumes from the firestorms created by deliberate incendiary bombing in Dresden and Hamburg, Germany, in World War II did not exceed 10 kilometers in height. The injection of smoke into the stratosphere (that is, above 15 kilometers, or 50,000 feet) seems highly unlikely.

Forest-fire expert Craig Chandler believes that the TTAPS researchers exaggerated their estimates of smoke from wildfires by a factor of two or more, because (1) for much of the year Northern Hemisphere forests are almost immune to ignition and (2) immediately upon a thermal-radiation pulse, forests emit steam, which helps shield them from catching on fire.

The TTAPS team's estimate probably additionally exaggerated by a factor of 10 the amount of smoke produced by a nuclear conflict, according to some nuclear-weapons-effects experts (such as Cresson Kearny from Oak Ridge National Laboratory) and the Defense Nuclear Agency. In itself, this lower level of smoke production would reduce the climatic effect from catastrophic to noticeable but not severe, because as optical density (smoke) declines, the penetration of light and thermal radiation increases very rapidly (exponentially).

Even if the TTAPS baseline calculation of smoke production were correct, it is of the same magnitude as current annual global smoke emission, which does not have catastrophic effects. The TTAPS team's claim, without evidence, that "nuclear smoke" is 100 times more effective in disturbing the atmosphere is not convincing.

Once dust and smoke particles are in the troposphere (the portion of the atmosphere below the stratosphere), they may be removed by coalescence (sticking together to form heavier particles that fall out faster), by rainfall, and by all other natural processes that currently scavenge out of the sky particles from fires, dust storms, smoke stacks, etc. The TTAPS researchers, however, underestimated these processes and consequently exaggerated the nuclear cloud's capacity to block sunlight (opacity).

Furthermore, in the presence of cooling, dust and smoke particles themselves act as condensation centers for the formation of raindrops and snowflakes, causing precipitation that washes out much of the remaining particles, which would further decrease the nuclear-winter effects. Studies performed at the Lawrence Livermore Laboratory in California, a major center for weapons research, suggest that this process would very effectively remove smoke from even the stratosphere, which is above the atmospheric level at which the normal "scavenging" of weather activity occurs. Physicist Joyce Penner, head of Livermore's nuclear-smoke research, simulated on a computer the exploding of a one-megaton bomb. The simulation showed that though smoke rose into the stratosphere, half rapidly fell back into the troposphere (where normal weather activity would wash it out quickly). Penner estimated that the half remaining in the stratosphere had almost three times the condensation needed to make rain. So even stratospheric smoke would be washed back to earth by storms of its own creation.

In fact, the TTAPS study did not take into account many of the moderating effects of water in atmospheric dynamics and radiative effects. The researchers appear to have ignored not only the generation of large amounts of water through combustion from the postulated fires, but also the much-larger mass of water that is naturally present in the atmosphere. Through continuous evaporation and precipitation, this water completely recycles itself in about one week. In their baseline case, the TTAPS researchers calculate the emission of 225 million tons of smoke, but this is about 10,000 times less than the mass of atmospheric water. Thus, according to nuclear physicist Edward Teller, the TTAPS assumption that dust and smoke particles would remain in the troposphere for weeks to months is highly suspect.

In their projections of atmospheric cooling, Turco and team also neglected the critical impact of the oceans, which cover most of Earth's surface. The sea would powerfully mitigate cooling from a nuclear cloud, because it has an immense capacity to hold heat and, at the surface, transfers heat to the atmosphere very well. Thus, no matter what the initial cooling over land surfaces, the ocean temperature will not change significantly, and it will act as a massive heat source to warm the atmosphere, rapidly increasing the temperature at the coastlines and eventually the interior temperatures, as well. Moreover, temperature differences at the shore would likely cause massive storms, which would accelerate the moderating and scavenging processes.

If a nuclear conflict produced 10 times less smoke than TTAPS predicted, and if other moderating processes (rain, for example) were to reduce by three times the average cloud opacities that TTAPS predicted, climatic effects would hardly be noticeable, except in the local areas of the explosions. The biological consequences of a nuclear cloud, too, would be negligible compared to the effects of the blast, fire, radiation, and fallout. As John Maddox stated eloquently in an editorial in the British science journal Nature, the nuclear winter hypothesis is not yet established, and "talk of some of the consequences of nuclear warfare had better be postponed until the underlying assumptions are better understood."

We do know, however, that nature has already "performed" an experiment that tends to refute the TTAPS predictions. In 1815, the largest and deadliest volcanic eruption in recorded history occurred with the explosion of Mount Tambora on the Indonesian island Sumbawa. Writing in Science magazine last year, NASA scientist R.B. Stothers calculated that this explosion injected approximately 200 million tons of particles into the stratosphere, on the same order predicted by the TTAPS baseline scenario. "Mean temperature in the Northern Hemisphere apparently dropped by 0.4 to 0.7 degree in 1816," Stothers further reported. There were remarkable meteorological phenomena, Stothers noted, but no long-lasting widespread winterlike effects or biological disasters.

In order to disarm critics who believe that their baseline case is not the only way a nuclear war might be fought, the TTAPS authors performed similar calculations on at least 18 different nuclear-exchange scenarios, with the "smallest" being the previously mentioned 100-megaton, air-burst attack on cities only. The researchers varied their assumptions of total explosive yield, warhead sizes, and percentage of yield hitting on cities, rural areas, etc. They concluded that many different scenarios above the 100-MT "threshold" would cause nuclear winter. Thus Paul Ehrlich states that "the [TTAPS] predictions of climatic changes are quite robust."

If this is true, why must Ehrlich and company use the worst-case scenario for their projections of biological effects (a 10,000-megaton exchange in which the properties of the cloud's particles "are assigned adverse but not implausible values and in which 30 percent of the soot is carried by firestorms to stratospheric altitudes")? Ehrlich and some of his colleagues seem to have a "strange love" for apocalyptic predictions.

In fact, even if one could accept the TTAPS researchers' methods, their nuclear-winter predictions are not robust. All of the cases they postulate to cause climatic changes involve substantial burning of cities, which contributes the bulk of smoke. At one place in their study, the TTAPS authors implicitly admit this. Yet Sagan and Ehrlich have not mentioned this assumption in their public presentations on the issue.

The National Research Council's December 1984 study of the nuclear-winter hypothesis—which TTAPS researcher Richard Turco says "legitimizes" the hypothesis—similarly overemphasizes attacks on cities in its nuclear-exchange scenario. The study assumed a 6,500-megaton war with the largest NATO and Warsaw Pact cities among the targets. The NRC calculations concluded that within only days following such a war a nuclear cloud could block 99 percent of the sun's light from reaching the Northern Hemisphere.

If urban areas—population centers, that is—are not deliberately attacked, nuclear winter is unlikely. If they are deliberately avoided, climatic catastrophe is probably impossible.

The suggestion that cities will not be burned in a nuclear war may seem surprising to anyone who saw the ABC television movie The Day After or the film Last Epidemic, presented by a group calling itself "Physicians for Social Responsibility." In the standard PSR presentation of a nuclear-war scenario, every major American population center is the direct target of massive air-burst attacks. These scenarios violate everything that is publicly known about Soviet targeting policy and are most probably contradictory to current US target priorities, as well.

Common sense dictates that in case of nuclear conflict, the first-priority targets must be the enemy's nuclear-weapons-launching facilities, including missile silos, submarine bases, and bomber bases. Of second priority would be command, control, and communications facilities, and third would be other military installations and non-nuclear forces. Fourth, perhaps, would be industrial targets that would provide an enemy military support in case of a long war. Populations per se should be last on the list—unless the goal is pure revenge or the civilian and military leadership have jointly gone insane with a desire for mass murder.

But there is more than common sense behind this assessment. Soviet military leaders have stated openly and repeatedly that their objective is not to turn the large economic and industrial regions into a heap of ruins but to destroy strategic combat means and paralyze military production. In other words, their basic strategy is one of counterforce—targeting enemy facilities of strategic military value—not countervalue (attacks on population and nonmilitary production facilities).

Much of the confusion and misunderstanding about this issue has been caused by widespread publicity for the idea of "mutual assured destruction" (MAD), a doctrine usually attributed to former Defense Secretary Robert McNamara. In case of massive attack on the United States and its allies, the doctrine asserted, enough nuclear weapons would be left—and would be used—to assure destruction of all major cities in the attacking country.

Even if this doctrine were ever desirable because of technological limitations that prevented weapons from accurately hitting only military targets, or for establishing mutual deterrence through a "balance of terror," subsequent technical developments have made weapons much more accurate for counterforce targeting—and the side-effects of even a pure counterforce attack are quite terrible enough to assure deterrence with current arsenals. President Reagan recently issued a directive on counterforce targeting—population centers per se are not to be attacked deliberately—and many experts now believe that US targeting strategy is now as "humane" as that of the USSR.

Nevertheless, many military-industrial targets are in or near cities, and even a "pure" counterforce attack would cause terrible collateral damage to civilian populations, especially if explosions are inaccurate or if especially powerful bombs are exploded in air bursts. Currently both the US and Soviet arsenals are being rapidly modernized, and the USSR—like the United States—is believed to be decreasing the explosive power of warheads as targeting becomes more accurate.

If the nuclear-winter prediction is scientifically valid at all, the maximum risk of its occurrence was probably in the late '50s and early '60s, during the era of multimegaton warheads, inaccurate missiles, and the MAD doctrine. The more recent increases in the number of tactical weapons and the advent of smaller but more accurate delivery systems (MIRVs and Cruise missiles) may have increased the total number of weapons but probably did not increase the risk of burning cities. In these days of "nuclear freeze" propaganda, few people realize that the total explosive yield of the US weapon stockpile is now only approximately a quarter of what it was two decades ago.

To sum up the discussion to this point: the TTAPS researchers' model calculations and assumptions are equivalent to a worst case. Although plausible as an extreme limit under certain nuclear-exchange scenarios, a worldwide, catastrophic nuclear winter is an unlikely outcome. Furthermore, the nuclear-winter hypothesis is much less plausible under scenarios that involve less city-burning than the TTAPS team generally assumed. And since common sense, current strategy, and technical trends all make countervalue (city-directed) attacks less likely, climatic disaster is rather improbable.

Yet despite the widely acknowledged uncertainties about the nuclear-winter hypothesis, many who ought to know better are presenting the prediction as all but certain. At the October-November 1983 nuclear-winter conference in Washington, for instance, conference spokesman Paul Ehrlich uttered a string of doomful pronouncements, as if only minor details remained to be worked out. "If there is a full-scale nuclear war," Ehrlich said, "odds are you can kiss the Northern Hemisphere good-bye." And: "There is simply no question of what the long-term effects of any size nuclear war are going to be. Catastrophic. So there is no need to hang tough, wait for more research, see what happens."

Carl Sagan, too, has promoted the hypothesis as if it is all but fact. In a Parade magazine article that appeared nearly one year after his original "Nuclear Winter" article of October 1983, for example, Sagan viewed the nuclear-winter hypothesis as sufficiently confirmed to render any civil defense "a dismal and largely untenable prospect." And when the National Research Council issued its nuclear-winter study in December 1984, Sagan quickly seized its findings as an endorsement of the TTAPS conclusions. Physicist Jonathan Katz, a member of the NRC study panel, corrected Sagan, noting in a letter to the New York Times that the NRC study "emphasized the very large uncertainties plaguing all calculations of this phenomenon." Katz concluded: "This difficult problem needs further research. It does not need partisans claiming that it is settled."

Unless it is proven impossible, the very idea of a nuclear winter makes it necessary to do more accurate calculations based on better models, with more realistic assumptions, more likely scenarios, and improved data about the emission of smoke, its properties of light absorption, etc. Various governmental bodies have now commissioned a number of studies to address what the December 1984 National Research Council study called "enormous uncertainties" still surrounding the nuclear-winter hypothesis. Among those performing or commissioning studies are the National Oceanic and Atmospheric Administration, the Defense Nuclear Agency, the Department of Energy, the National Aeronautics and Space Administration, and the Environmental Protection Agency.

If the plausibility of the nuclear-winter hypothesis is to be accurately determined, such studies must, for example, develop a realistic model of how dust and smoke would spread throughout the atmosphere both longitudinally and latitudinally. Just as essential are models for the distribution of smoke and dust over time, assuming more realistically than the TTAPS study does that all nuclear attacks would not occur simultaneously but spasmodically.

How water contributes to the scavenging of smoke particles and how the oceans mitigate temperature changes also are important to investigate. Moreover, all data from previous atmospheric weapons tests and from World II firebombing should be released, so that this information can be reviewed to refine our knowledge of how dust, smoke, and soot are injected into the atmosphere.

A nuclear winter, if it occurred, would be felt far from the site of attack and thus redound to the detriment of the attacker. It could affect every nation in the Northern Hemisphere (effects in the Southern Hemisphere are expected to be much less). Thus, data and calculations on the nuclear-winter hypothesis should be shared with the USSR and all nuclear-armed nations, as it would be in all of our interests to prevent such an occurrence. This is one hypothesis that no one wants to test with a full-scale experiment.

If more-refined calculations still suggest the possibility of a nuclear winter, but with uncertainty due to lack of sufficient data about dust or smoke injection from actual tests, we might consider a minimal atmospheric test series, especially near a forest or simulated city, done in cooperation with the USSR, with other nations as observers. This sounds outrageous only until one realizes that "the fate of the earth" could hang in the balance. It would be an absurd irony if the atmospheric test ban treaty (signed by the United States, the USSR, and other nations in 1962) has "backfired" on the world, leading to the buildup of potentially suicidal nuclear stockpiles because atmospheric effects of nuclear war are not understood. (Obviously, any such tests should be done with every effort to avoid sending radioactive fallout toward downwind populations.)

A "freeze" on present nuclear stockpiles and weapons development, however, would be an especially absurd response to the fear of nuclear winter: it would keep in place the very weapons that are alleged to be capable of causing global catastrophe—very powerful weapons that are less accurate than weapons now being developed. A much more logical response would be to negotiate a "build-down" (for example, eliminating two old weapons for each new one produced), or mutual "deep cuts" (for example, each side eliminating 20 or 30 percent of the total megatonnage of their arsenals). Most vital is to eliminate the weapons with explosive yields greater than one megaton, which produce the most collateral and environmental damage, the largest area of fire in air bursts, and the largest amount of stratospheric dust from surface bursts. These have largely been removed from the US stockpile but are still somewhat prevalent in the USSR arsenal. If the nuclear-winter hypothesis has any validity, it is in our interest (and the Soviets' as well) to do away with these weapons. And, of course, there are moral and environmental reasons for doing away with these weapons, as well.

Strategic-arms reduction and modernization are also desirable on these grounds. But a "freeze now" could make it more likely that we will all freeze later in a nuclear winter.

As I noted earlier, calculations of the likelihood of nuclear winter strongly depend on assumptions that 100–1,000 cities will burn. So if the nuclear-winter prediction has any validity, it would imply that urban areas should be removed from target priority lists on practical grounds of self-interest, as well as on moral grounds. In his 1962 book Thinking About the Unthinkable, Herman Kahn anticipated this strategy when he detailed a "counterforce plus avoidance" targeting policy: only militarily strategic sites should be attacked, with deliberate efforts to spare people and civilian property. Indeed, the possibility of a nuclear winter makes the MAD doctrine look more insane than ever and is another argument for counterforce instead of countervalue military strategy.

At conferences on the nuclear-winter idea in late 1983, publicists of the idea proudly announced that this new argument would "put the final nail in the coffin" of civil defense. Paraphrasing Mark Twain, rumors of the death of civil defense are again highly exaggerated.

It may seem paradoxical, but if the chance of a nuclear winter is justified scientifically and accepted by the nuclear powers, long-term civil defense and medical preparations will be all the more desirable. If military planners appreciate that a city-burning strategy might cause mutual suicide from a nuclear winter, they will continue to shift toward a "counterforce plus avoidance" strategy. This would mean fewer fire-producing air bursts over cities and more fallout-producing ground bursts on military targets, with more people surviving attack but needing shelter from increased radioactive fallout, as well as food, water, medical attention, and supplies. Civil defense is still necessary for future scenarios of all-out nuclear war, as well as for all other possibilities (accidental detonation, limited war, terrorist explosion, attack by a smaller nation with a crazed dictator, etc.).

A portion of current expenditures on "defense" (offense, in reality) should be shifted into truly passive (that is, civil) defense. A shelter cannot directly kill a Russian or any other person, while both old and new weapons can. Lest the USSR misunderstand this action as a prelude to attack, we should invite the Soviets to join in an open race to protect both our populations instead of making them more vulnerable to mutual destruction. Even if strategic-arms-reduction talks are successful, there will be periods of danger during reduction when war may be more likely because of suspicions about mutuality, worries about inadequate verification, doubts about the adequacy of a remaining deterrent, etc. Adequate civil defense could help to bridge this credibility gap on the way to mutual assured survival.

Nuclear winter may be possible, but it is highly improbable. To the extent that it is possible, the risk can be reduced by shifting strategy from targeting attacks on cities to "counterforce plus avoidance" and by emphasizing civil defense and preparedness, along with bilateral strategic-arms reduction and modernization of weapons instead of a nuclear freeze.

Apocalyptic predictions of the end of humanity, however, in spreading the message that such measures are futile, can only worsen our present unenviable situation. Such predictions are, as usual, more useful for irresponsible propaganda than for the kind of careful thinking that is necessary to avoid nuclear disaster while preserving freedom from potential aggressors.

Howard Maccabee is a physician whose specialty is radiation therapy for cancer patients. He also holds a Ph.D. from the University of California at Berkeley, earned jointly from the departments of nuclear engineering and medical physics, with dissertation research in radiation biophysics. From 1982 to 1984, he was president of Doctors for Disaster Preparedness.