Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
The Medical Implications of Nuclear War, Institute of Medicine. @) 1986 by the National Academy of Sciences. National Academy Press, Washington, D.C. Nuclear Winter: The State of the Science GEORGE F. CARRIER, PH.D. Harvard University, Cambridge, Massachusetts During the past year, it has become widely known that a major exchange of nuclear weapons could result in, among other consequences, a signif- icant contamination of a large portion of the earth's atmosphere (NRC, 1985; Crutzen and Birks, 1982; Turco et al., 1983; and Sagan, 1983- 1984~. This contamination, preliminary calculations have suggested, could lead to cooling of significant portions of the earth's surface a "nuclear winter." There is little doubt that atmospheric modifications of this char- acter would occur. But their extent and duration and hence their potential impact on people, food supplies, and other biological systems are very difficult to determine, and they remain controversial. In the following, I describe the principal types of contamination and the uncertainties aKen- dant upon calculations of the atmospheric effects, given our present, lim- ited knowledge. The fireballs caused by nuclear weapons directed against hardened m~- itary targets and therefore detonated at ground level would contain large numbers of dust particles in the submicron size that is, with typical dimensions of less than one ten-thousandth of a centimeter as well as large amounts of nitrogen oxides (NOx). A major portion of both of these substances would be earned into the stratosphere where, in otherwise unmodified circumstances, most would remain for a considerable period (on the order of one year) while being gradually removed by natural processes. The nitrogen oxides would deplete stratospheric ozone and increase the flux of ultraviolet radiation reaching the lower atmosphere; the dust would reduce somewhat the total amount of sunlight reaching the lower atmosphere. 136
NUCLEAR WINIER: THE STATE OF THE SCIENCE 137 Of greater concern are potential modifications to the lower atmosphere. Weapons directed at targets in or close to cities and detonated in the air would ignite intense, extensive fires. The fires in turn would generate large numbers of smoke particles also of submicron size. The smoke would rise to moderately high altitudes (four to nine kilometers), where it could impede the passage of sunlight and alter many of the details of the heat balance and motions of the atmosphere. In particular, published reports say, temperatures near the ground could be significantly lowered. It is also possible that modifications of the winds at high altitudes could delay or speed the removal of NOx and dust from the stratosphere. Any assessment of this potential threat depends on quantitative estimates of the numbers of weapons that might be used against the various types of targets, the yields of those weapons, the amounts of contaminants (dust, NOx, and smoke) that would be produced, and their lateral and vertical distribution in the atmosphere. The assessment also depends on calcula- tions of the atmosphere's response to the presence of those contaminants that is, the evolving temperature distributions and motions. However, any attempt to make such calculations with today's knowledge and today's understanding of many of the pertinent phenomena is severely impeded by a large number of major uncertainties. To understand the extent of those uncertainties and their role in attempts to estimate the degree of the atmospheric degradation that would follow a nuclear war, it may be useful to consider the ways in which uncertainties would be compounded in the events that accompany a major weapons exchange. There are three types of uncertainties. These concern the nuclear weapons scenario, the production of smoke and its injection into the atmosphere, and the atmospheric response to contaminants on the scale · · . envisioned. The first set of uncertainties cannot be removed. One cannot know in advance of the nuclear phase of the postulated hostilities, for example, the numbers of weapons that any combatant would actually use, the dis- tributions of targets against which those weapons would be directed, or the number of those weapons that would reach their targets and detonate successfully. One can postulate, however, a plausible hypothetical ex- change and the time of year at which it is to occur and then try to estimate the atmospheric degradation caused by that exchange. In contrast, the second set of uncertainties can be estimated by a process illustrated in the following example. A moderate amount of observational data exists concerning large fires in irregularly littered solid fuel, such as would be found in a city in the aftermath of a nuclear explosion (McMahon, 1983~. These data suggest that between 2 and 6 percent of the fuel actually burned would be converted to smoke. The data do not imply that the
138 PHYSICAL EFFECTS AND ENVIRONMENTAL CONSEQUENCES fraction converted to smoke cannot be larger; in fact, if the fuel largely consisted of synthetic organic materials, it is known that the smoke pro- duction could be much larger than 6 percent. Alternatively, distributions of fuel and air supply are possible for which smoke production can be much lower than 6 percent. Nevertheless, no competing arguments seem to refute the 2 to 6 percent range, which we will refer to as the uncertainty range. Furthermore, because the largest number in this range is three times the smallest, we will say that the uncertainty factor is three. The size of the smoke particles and the height to which they rise in the atmosphere are important because a given mass of larger particles will impede the passage of solar radiation less effectively than will the same mass of smaller particles. Furthermore, larger particles and those injected at lower altitudes will be more rapidly removed. To estimate the amount of submicron smoke that would rise above this altitude requires quantitative estimates for the amount of fuel in the regions where burning would occur (the fuel supply), the fraction of the fuel supply that would burn, and the fraction of the fuel burned that would emerge as smoke. It also requires estimates of the fraction of smoke particles that would remain at submicron size during their ascent in the smoke plume, despite their coagulation and incorporation into moisture condensation droplets Mat would form at the higher altitudes. I would assert that the uncertainty factor for the fuel supply is not less than two, that the uncertainty factor in the fraction burned is not less than two, that the uncertainty factor in the fraction of fuel burned that becomes smoke is not less than three, and that the un- certainty factor in the nonagglomerated fraction of the total smoke is not less than three. Thus, the composite uncertainty factor associated with this second set of uncertainties is not less than 36. Still other uncertainties are not included in this estimate: the height of the smoke plume (and hence the height at which the smoke is injected); the optical properties of the smoke (the more opaque the smoke, the more it obscures sunlight); and changes in the smoke's optical properties over a period of time. In the National Research Council's recent report, The Effects on the Atmosphere of a Major Nuclear Exchange, a particular scenario for a nuclear exchange in which somewhat less than half (6,500 megatons) of the world's arsenal is expended was adopted as a baseline case. In other words, this scenario was used to illustrate the process of estimating the atmospheric effects of a nuclear exchange. No pretense is made that this is a "most likely exchange." It is merely a plausible assumption whose estimated consequences can give some guidance regarding possible at- mospheric degradation. For this assumed nuclear exchange, the amount of submicron smoke that would survive the ascent in the fire plume is between 20 million tons and 650 million tons. These numbers are generally
NUCLEAR WINTER: THE STATE OF THE SCIENCE 139 consistent with the uncertainty factors given above. (Some small and unimportant discrepancies arise, however, because this discussion is a highly simplified recasting of the National Research Council's report.) In that report, for purposes of inquiry, the investigators chose to assume that 180 million tons of submicron smoke were injected at altitude (four to nine kilometers) in the atmosphere. The third set of uncertainties those dealing with the atmosphere's response complicates the final stage of analysis. Atmospheric scientists have at their disposal a variety of computational procedures designed to reproduce some of the large-scale features of the atmosphere's response to various conditions. These mathematical models are designed to deal with relatively small variations in normal atmospheric behavior. The mod- eling of small-scale processes (such as precipitation, particle removal, the mixing effects of turbulence, to name a few) are chosen and refined so that they satisfactorily represent the large-scale consequences of those small-scale processes. They are satisfactory because they are designed, insofar as possible, to conform to the observed behavior of the normal atmosphere. In the phenomena of interest here, however, the conditions include strong and abnormal temperature gradients and millions of tons of smoke particles at an altitude of several kilometers, yet there are no observations of an atmosphere in such a severely modified state that could be used to validate the models. Accordingly, it is especially difficult to assess quantitatively the inaccuracies that may result when making cal- culations with existing mathematical models. Clearly, it is eminently sen- sible to use these models to estimate the order of magnitude of the temperature change caused by smoke, but the results can only be regarded as sugges- tive. They are definitely not predictions. A variety of computational models have been applied to the baseline war scenario described above and to some variations on that case (NRC, 1985; Crutzen and Birks, 1982; Turco et al., 1983; McCracken, 1983; Thompson et al., 1984~. The results must be interpreted with care, but they boil down to the suggestion that the atmospheric response to smoke injection on the order of 180 million tons, as estimated using currently available computational models, would include temperature changes that could be of serious concern. In particular, the results suggest that for an exchange occurring in the summer, with all of the foregoing quantitative uncertainty, intermittent temperature drops in the northern temperate zone could be on the order of 20 degrees centigrade and might continue for a few weeks. Although it is even more uncertain, smaller temperature drops might occur in the tropics of the northern hemisphere. It is even possible that areas in the southern hemisphere could experience longlasting tem- perature drops of several degrees.
140 PHYSICAL EFFECTS AND ENVIRONMENTAL CONSEQUENCES From this discussion and the studies on which it is based, I find un- avoidable the following three-part conclusion: 1. The uncertainties that pervade the quantitative assessment of He atmospheric effects of a major nuclear exchange are so numerous and so large that no definitive description of those effects is possible at this time. Nevertheless: 2. Ihe model calculations Mat can be made suggest temperature changes of a size that could have very severe consequences. This possibility cannot and must not be ignored. Therefore: 3. It is incumbent on agencies having resources that can be allocated to such matters and on appropriate members of the scientific and tech- nological community to support and conduct investigations Tat can narrow many of the uncertainties. Only in this way can we approach a posture from which a more definitive assessment can be made. Subsequent to the appearance of the foregoing article, in Issues in Science and Technology (Winter 1985:114-117), the response of the atmosphere has been recalculated several times using models which should replicate some features of the real phenomena in a more realistic way. The results of these calculations do differ in some of their details from the earlier results but those differences and the uncertainties that remain are such that no changes in the conclusions cited above are justified. REFERENCES Crutzen, P. J., and J. W. Birks. 1982. The atmosphere after a nuclear war: Twilight at noon. Ambio 11:114-125. McCracken, M. 1983. Nuclear War: Preliminary Estimates of the Climatic Effects of a Nuclear Exchange. Paper presented at the Third Conference on Nuclear War, Erice, Sicily, August 12-23. McMahon, C. K. 1983. Characteristics of Forest Fuels, Fires, and Emissions. Paper pre- sented at the 76th Annual Meeting of the Air Pollution Control Association, Atlanta, Georgia, June 19-24. National Research Council. 1985. The Effects on the Atmosphere of a Major Nuclear Exchange. Washington, D.C.: National Academy Press. Sagan, C. 1983-1984. Foreign Affairs (Winter):257-292. Thompson, S. L., V. V. Aleksandrov, G. L. Gtenchikov, S. H. Schneider, C. Covey, and R. M. Chervin. 1984. Global consequences of nuclear war: Simulations with three dimensional models. Ambio 13(4):236-243. Turco, R. P., O. B. Toon, T. P. Ackerman, J. B. Pollock, and C. Sagan. 1983. Nuclear winter: Global consequences of multiple nuclear explosions. Science December 23: 1283- 1292.