Nuclear winter: global consequences of multple nuclear explosions

10 Pages·2.372 MB·English

by  Turco R.P., Toon O.B., et al.

23 December 1983, Volume 222, Number 4630
exchange and would inherit the postwar
environment. Accordingly, the longer-
term and global-scale aftereffects of nu­
clear war might prove to be as important
as the immediate consequences of the
Nuclear Winter: Global Consequences war.
To study these phenomena, we used a
of Multiple Nuclear Explosions series of physical models: a nuclear war
scenario model, a particle microphysics
model, and a radiative-convective mod­
R. P. Turco, O. B. Toon, T. P. Ackerman el. The nuclear war scenario model spec­
ifies the altitude-dependent dust, smoke,
J. B. Pollack, Carl Sagan
radioactivity, and NO* injections for
each explosion in a nuclear exchange
(assuming the size, number, and type of
detonations, including heights of burst,
Concern has been raised over the quantities of sooty smoke that would geographic locales, and fission yield
short- and long-term consequences of attenuate sunlight and perturb the cli­ fractions). The source model parameter­
the dust, smoke, radioactivity, and toxic mate. These developments have led us to ization is discussed below and in a more
vapors that would be generated by a calculate, using new data and improved detailed report (75). The one-dimension­
nuclear war (7-7). The discovery that models, the potential global environmen- al microphysical model (15-17) predicts
the temporal evolution of dust and
smoke clouds, which are taken to be
Summary. The potential global atmospheric and climatic consequences of nuclear rapidly and uniformly dispersed. The
war are investigated using models previously developed to study the effects of one-dimensional radiative-convective
volcanic eruptions. Although the results are necessarily imprecise, due to a wide model (1-D RCM) uses the calculated
range of possible scenarios and uncertainty in physical parameters, the most dust and smoke particle size distribu­
probable first-order effects are serious. Significant hemispherical attenuation of the tions and optical constants and Mie the­
solar radiation flux and subfreezing land temperatures may be caused by fine dust ory to calculate visible and infrared opti­
raised in high-yield nuclear surface bursts and by smoke from city and forest fires cal properties, light fluxes, and air tem­
ignited by airbursts of all yields. For many simulated exchanges of several thousand peratures as a function of time and
megatons, in which dust and smoke are generated and encircle the earth within 1 to 2 height. Because the calculated air tem­
weeks, average light levels can be reduced to a few percent of ambient and land peratures are sensitive to surface heat
temperatures can reach -15° to -25°C. The yield threshold for major optical and capacities, separate simulations are per­
climatic consequences may be very low: only about 100 megatons detonated over formed for land and ocean environ­
major urban centers can create average hemispheric smoke optical depths greater ments, to define possible temperature
than 2 for weeks and, even in summer, subfreezing land temperatures for months. In contrasts. The techniques used in our 1-
a 5000-megaton war, at northern mid-latitude sites remote from targets, radioactive D RCM calculations are well docu­
fallout on time scales of days to weeks can lead to chronic mean doses of up to 50 mented (75, 18).
rads from external whole-body gamma-ray exposure, with a likely equal or greater Although the models we used can pro­
internal dose from biologically active radionuclides. Large horizontal and vertical vide rough estimates of the average ef­
temperature gradients caused by absorption of sunlight in smoke and dust clouds fects of widespread dust and smoke
may greatly accelerate transport of particles and radioactivity from the Northern clouds, they cannot accurately forecast
Hemisphere to the Southern Hemisphere. When combined with the prompt destruc short-term or local effects. The applica­
tion from nuclear blast, fires, and fallout and the later enhancement of solar ultraviolet bility of our results depends on the rate
radiation due to ozone depletion, long-term exposure to cold, dark, and radioactivity and extent of dispersion of the explosion
could pose a serious threat to human survivors and to other species. clouds and fire plumes. Soon after a
large nuclear exchange, thousands of in­
dividual dust and smoke clouds would be
dense clouds of soil particles may have tal effects of dust and smoke clouds distributed throughout the northern mid-
played a major role in past mass extinc­ (henceforth referred to as nuclear dust latitudes and at altitudes up to 30 km.
tions of life on Earth (8-10) has encour­ and nuclear smoke) generated in a nucle­
aged the reconsideration of nuclear war ar war (27). We neglect the short-term R. P. Turco is at R & D Associates, Marina del
Rey, California 90291; O. B. Toon, T. P. Ackerman,
effects. Also, Crutzen and Birks (7) re­ effects of blast, fire, and radiation (12-
and J. B. Pollack are at NASA Ames Research
cently suggested that massive fires ignit­ 14). Most of the world's population Center, Moffett Field, California 94035; and Carl
Sagan is at Cornell University, Ithaca, New York
ed by nuclear explosions could generate could probably survive the initial nuclear 14853.
23 DECEMBER 1983 1283
Horizontal turbulent diffusion, vertical though the total number of high-yield The properties of nuclear dust and
wind shear, and continuing smoke emis­ warheads is declining with time, about smoke are critical to the present analy­
sion could spread the clouds of nuclear 7000 MT is still accounted for by war­ sis. The basic parameterizations are de­
debris over the entire zone, and tend to heads of > 1 MT. There are also scribed in Tables 2 and 3, respectively;
fill in any holes in the clouds, within 1 to - 30,000 lower-yield tactical warheads details may be found in (75). For each
2 weeks. Spatially averaged simulations and munitions which are ignored in this explosion scenario, the fundamental
of this initial period of cloud spreading analysis. Scenarios for the possible use quantities that must be known to make
must be viewed with caution; effects of nuclear weapons are complex and optical and climate predictions are the
would be smaller at some locations and controversial. Historically, studies of the total atmospheric injections of fine dust
larger at others, and would be highly long-term effects of nuclear war have (:s 10 |xm in radius) and soot.
variable with time at any given location. focused on a full-scale exchange in the Nuclear explosions at or near the
The present results also do not reflect range of 5000 to 10,000 MT (2, 72, 20). ground can generate fine particles by
the strong coupling between atmospheric Such exchanges are possible, given the several mechanisms (27): (i) ejection and
motions on all length scales and the current arsenals and the unpredictable disaggregation of soil particles (29), (ii)
modified atmospheric solar and infrared nature of warfare, particularly nuclear vaporization and renucleation of earth
heating and cooling rates computed with warfare, in which escalating massive ex­ and rock (30), and (iii) blowoflf and
the 1-D RCM. Global circulation pat­ changes could occur (25). sweepup of surface dust and smoke (57).
terns would almost certainly be altered An outline of the scenarios adopted Analyses of nuclear test data indicate
in response to the large disturbances in here is presented in Table 1. Our base­ that roughly 1 x 105 to 6 x 105 tons of
the driving forces calculated here (79). line scenario assumes an exchange of dust per megaton of explosive yield are
Although the 1-D RCM can predict only 5000 MT. Other cases span a range of held in the stabilized clouds of land sur­
horizontally, diurnally and seasonally total yield from 100 to 25,000 MT. Many face detonations (32). Moreover, size
?
averaged conditions, it is capable of esti­ high-priority military and industrial as­ analysis of dust samples collected in
mating the first-order climate responses sets are located near or within urban nuclear clouds indicates a substantial
of the atmosphere, which is our intention zones (26). Accordingly, a modest frac­ submicrometer fraction (33). Nuclear
in this study. tion (15 to 30 percent) of the total yield is surface detonations may be much more
assigned to urban or industrial targets. efficient in generating fine dust than vol­
Because of the large yields of strategic canic eruptions (75, 34), which have
Scenarios warheads [generally ^ 100 kilotons been used inappropriately in the past to
(KT)], "surgical" strikes against individ­ estimate the impacts of nuclear war (2).
A review of the world's nuclear arse­ ual targets are difficult; for instance, a The intense light emitted by a nuclear
nals (20-24) shows that the primary stra­ 100-KT airburst can level and burn an fireball is sufficient to ignite flammable
tegic arid theater weapons amount to area of - 50 km2, and a 1-MT airburst, materials over a wide area (27). The
=* 12,000 megatons (MT) of yield carried = 5 times that area (27, 28), implying explosions over Hiroshima and Nagasaki
by — 17,000 warheads. These arsenals widespread collateral damage in any both initiated massive conflagrations
are roughly equivalent in explosive pow­ "counter-value," and many "counter- (35). In each city, the region heavily
er to 1 million Hiroshima bombs. Al­ force/ ' detonations. damaged by blast was also consumed by
fire (36). Assessments over the past two
decades strongly suggest that wide­
spread fires would occur after most nu­
Table 1. Nuclear exchange scenarios.
clear bursts over forests and cities (37-
Percent of yield 44). The Northern Hemisphere has
= 4 x Ю7 km2 of forest land, which
Ur­
Total holds combustible material averaging
Case* Tyioetladl Sur­ boarn Wyairehldea d numofb er — 2.2 g/cm2 (7). The world's urban and
(MT) face indus­ range explo­ suburban zones cover an area of
bursts trial (MT) sions =-1.5 x Ю6 km2 (75). Central cities,
tar­
which occupy 5 to 10 percent of the total
gets
urban area, hold - 10 to 40 g/cm2 of
1. Baseline exchange 5,000 57 20 0.1 to 10 10,400 combustible material, while residential
2. Low-yield airbursts 5,000 10 33 0.1 to 1 22,500t areas hold - 1 to 5 g/cm2 (41, 42, 44,
9. 10,000-МТФ maximum 10,000 63 15 0.1 to 10 16,160
45). Smoke emissions from wildfires and
10. 3,000-MT exchange 3,000 50 25 , 0.3 to 5 5,433
large-scale urban fires probably lie in the
11. 3,000-MT counterforce 3,000 70 0 1 to 10 2,150
12. l,000-MTexchange§ 1,000 50 25 0.2 to 1 2,250 range of 2 to 8 percent by mass of the
13. 300-MT Southern 300 0 50 1. 300 fuel burned (46). The highly absorbing
Hemisphere II
sooty fraction (principally graphitic car­
14. 100-MT city attackf 100 0 100 0.1 1,000
bon) could comprise up to 50 percent of
16. Silos, "severe" case# 5,000 100 0 5 to 10 700
18. 25,000-МТФ "future war" 25,000 72 10 0.1 to 10 28,300t the emission by weight (47, 48). In wild­
fires, and probably urban fires, ^ 90
♦Case numbers correspond to a complete list given in (15). Detailed detonation inventories are not
reproduced here. Except as noted, attacks are concentrated in the NH. Baseline dust and smoke parameters percent of the smoke mass consists of
are described in Tables 2 and 3. tAssumes more extensive MIRVing of existing missiles and some particles < 1 u,m in radius (49). For
possible new deployment of medium- and long-range missiles (20-23). {Although these larger total yields
might imply involvement of the entire globe in the war, for ease of comparison hemispherically averaged calculations at visible wavelengths,
results are still considered. §Nominal area of wildfires is reduced from 5 x 10s to 5 x 104
km2. 11 Nominal area of wildfires is reduced from 5 x 105 to 5 x 103 km2. HThe central city burden of smoke particles are assigned an imagi­
combustibles is 20 g/cm2 (twice that in the baseline case) and the net fire smoke emission is 0.026 g per gram nary part of the refractive index of 0.3
of material burned. There is a negligible contribution to the opacity from wildfires and dust. #Includes a
sixfold increase in the fine dust mass lofted per megaton of yield. (50).
1284 SCIENCE, VOL. 222
Simulations with an average surface temperature de­ smoke emission factor is 0.026 gram of
crease of ts 1 К (57-60). smoke per gram of material burned (ver­
The model predictions discussed here Case 14 represents a 100-MT attack on sus the conservative figure of 0.011 g/g
generally represent effects averaged over cities with 1000 100-KT warheads. In the adopted for central city fires in the base­
the Northern Hemisphere (NH). The ini­ attack, 25,000 km2 of built-up urban area line case). About 130 million tons of
tial nuclear explosions and fires would is burned (such an area could be ac­ urban smoke is injected into the tropo­
be largely confined (51) to northern mid- counted for by — 100 major cities). The sphere in each case (none reaches the
latitudes (30° to 60°N). Accordingly, the smoke emission is computed with fire stratosphere in case 14). In the baseline
predicted mean dust and smoke opacity parameters that differ from the baseline case, only about 10 percent of the urban
could be larger by a factor of 2 to 3 at case. The average burden of combustible smoke originates from fires in city cen­
mid-latitudes, but smaller elsewhere. material in city centers is 20 g/cm2 (ver­ ters (Table 3).
Hemispherically averaged optical depths sus 10 g/cm2 in case 1) and the average The smoke injection threshold for ma-
at visible wavelengths (52) for the mixed
nuclear dust and smoke clouds corre­
Table 2. Dust parameterization for the baseline case.
sponding to the scenarios in Table 1 are
shown in Fig. 1. The vertical optical
Materials in stabilized nuclear explosion clouds*
depth is a convenient diagnostic of nu­
~ Dust size
л H0
clear cloud properties and may be used Type of burst Dust mass distribution! 2
(ton/MT): (ton/MT):
roughly to scale atmospheric light levels [rm(|xm)/o7a]:
and temperatures for the various scenari­
os. Land surface 3.3 x 105 0.25/2.0/4.0 1.0 x 105
Land near-surface 1.0 x 105 0.25/2.0/4.0 1.0 x 105
In the baseline scenario (case 1, 5000
MT), the initial NH optical depth is = 4, Dust composition: siliceous minerals and glasses
of which - 1 is due to stratospheric dust Index of refraction at visible wavelengths^: n = 1.50 - 0.001 /
and = 3 to tropospheric smoke. After 1 Stabilized nuclear cloud top and bottom heights, zt and Zb, for surface and low-air bursts§:
z = 21 Y°2; z = 13 Y° 2; where Y = yield in megatons
month the optical depth is still — 2. t b
Multiburst interactions are ignored
Beyond 2 to 3 months, dust dominates
Baseline dust injections
the optical effects, as the soot is largely
Total dust — 9.6 x Ю8 tons; 80 percent in the stratosphere; 8.4 percent < 1 u,m in radius
depleted by rainout and washout (54). In
Submicrometer dust injection is ~ 25 ton/KT for surface bursts, which represents — 0.5 percent
the baseline case, about 240,000 km2 of of the total ejecta mass
urban area is partially (50 percent) Total initial area of stabilized fireballs — 2.0 x Ю6 km2
burned by - 1000 MT of explosions
♦Materials are assumed to be uniformly distributed in the clouds. tParticle size distributions (number/
(only 20 percent of the total exchange cnmor3m -a l |xnmum rbaedriu ms)o adree rlaodgi-unso ramnda ls iwzeit hv aar ipaonwcee,r -rleaswp etcatiilv aetl yl,a ragned saiz iess .t hTeh eex ppaornaemnte toefr st hrem ra~na dd ae paerned tehnec el oagt-
yield). This roughly corresponds to one large sizes. The log-normal and power-law distributions are connected at a radius of = 1 jim (15). $The
sixth of the world's urbanized land area, refractive indices of dust at infrared wavelengths are discussed in (10). §The model of Foley and
Ruderman (87) is adopted, but with the cloud heights lowered by about 0.5 km. The original cloud heights are
one fourth of the developed area of the based on U.S. Pacific test data, and may overestimate the heights at mid-latitudes by several kilometers.
NH, and one half of the area of urban
centers with populations > 100,000 in
the NATO and Warsaw Pact countries. Table 3. Fire and smoke parameterization for the baseline case.
The mean quantity of combustible mate­
Fire area and emissions
rial consumed over the burned area is Area of urban fire ignition defined by the 20 cal/cm2 thermal irradiance contour (= 5 psi peak
— 1.9 g/cm2. Wildfires ignited by the overpressure contour) with an average atmospheric transmittance of 50 percent:
remaining 4000 MT of yield burn another A (km2) = 250 Y, where Y = yield in megatons detonated over cities; overlap of fire zones is
ignored
500,000 km2 of forest, brush, and grass­ Urban flammable material burdens average 3 g/cm2 in suburban areas and 10 g/cm2 in city
lands (7, 39, 55), consuming — 0.5 g/cm2 centers (5 percent of the total urban area)
of fuel in the process (7). Average consumption of flammables in urban fires is 1.9 g/cm2
Total smoke emission in the baseline Average net smoke emission factor is 0.027 g per gram of material burned (for urban centers it is
only 0.011 g/g)
case is — 225 million tons (released over Area of wildfires is 5 x 105 km2 with 0.5 g/cm2 of fuel burned, and a smoke emission factor of
several days). By comparison, the cur­ 0.032 g/g
rent annual global smoke emission is Long-term fires burn 3 x 1014 g of fuel with an emission factor of 0.05 g/g
estimated as — 200 million tons (15), but Fire plume heights (top and bottom altitudes)
is probably < 1 percent as effective as Urban fires: 1 to 7 km
nuclear smoke would be in perturbing Fire storms (5 percent of urban fires): Zb ^ 5 km; zt ^ 19 km
Wildfires: 1 to 5 km
the atmosphere (56).
Long-term fires: 0 to 2 km
The optical depth simulations for cas­
Fire duration
es 1, 2, 9, and 10 in Fig. 1 show that a
Urban fires, 1 day; wildfires, 10 days; long-terrh fires, 30 days
range of exchanges between 3000 and
Smoke properties
10,000 MT might create similar effects.
Density, 1.0 g/cm3; complex index of refraction, 1.75 - 0.30 i; size distribution, log-normal
Even cases 11, 12, and 13, while less
with r (u,m)/a = 0.1/2.0 for urban fires and 0.05/2.0 for wildfires and long-term fires
m
severe in their absolute impact, produce
Baseline smoke injections
optical depths comparable to or exceed­ Total smoke emission = 2.25 x Ю8 tons, 5 percent in the stratosphere
ing those of a major volcanic eruption. It Urban-suburban fires account for 52 percent of emissions, fire storms for 7 percent, wildfires
is noteworthy that eruptions such as for 34 percent, and long-term fires for 7 percent
Tambora in 1815 may have produced Total area burned by urban-suburban fires is 2.3 x Ю5 km2; fire storms, 1.2 x Ю4 km2; and
wildfires, 5.0 x 105 km2
significant climate perturbations, even
23 DECEMBER 1983 1285
jor optical perturbations on a hemispher­ tures, with a minimum again near 250 K. land temperature decreases, particularly
ic scale appears to lie at - 1 x 108 tons. The temperature recovery in this in­ in coastal regions (10). The effect is
From case 14, one can envision the re­ stance is hastened by the absorption of difficult to assess because disturbances
lease of — 1 x Ю6 tons of smoke from sunlight in optically thin remnant soot in atmospheric circulation patterns are
each of 100 major city fires consuming clouds (see below). Comparable ex­ likely. Actual temperature decreases in
= 4 x 107 tons of combustible material changes with and without smoke emis­ continental interiors might be roughly 30
per city. Such fires could be ignited by sion (for instance, cases 10 and 11) show percent smaller than predicted here, and
100 MT of nuclear explosions. Unex­ that the tropospheric soot layers cause a along coastlines 70 percent smaller (10).
pectedly, less than 1 percent of the exist­ sudden surface cooling of short duration, In the baseline case, therefore, continen­
ing strategic arsenals, if targeted on cit­ while fine stratospheric dust is responsi­ tal temperatures may fall to = 260 К
ies, could produce optical (and climatic) ble for prolonged cooling lasting a year before returning to ambient.
disturbances much larger than those pre­ or more. [Climatologically, a long-term Predicted changes in the vertical tem­
viously associated with a massive nucle­ surface cooling of only \°C is significant perature profile for the baseline nuclear
ar exchange of - 10,000 MT (2). (60).] In all instances, nuclear dust acts exchange are illustrated as a function of
Figure 2 shows the surface tempera­ to cool the earth's surface; soot also time in Fig. 3. The dominant features of
ture perturbation over continental land tends to cool the surface except when the temperature perturbation are a large
areas in the NH calculated from the dust the soot cloud is both optically thin and warming (up to 80 K) of the lower strato­
and smoke optical depths for several located near the surface [an unimportant sphere and upper troposphere, and a
scenarios» Most striking are the extreme­ case because only relatively small tran­ large cooling (up to 40 K) of the surface
ly low temperatures occurring within 3 to sient warmings ^ 2 К can thereby be and lower troposphere. The warming is
4 weeks after a major exchange. In the achieved (67)]. caused by absorption of solar radiation
baseline 5000-MT case, a minimum land Predicted air temperature variations in the upper-level dust and smoke
temperature of =- 250 К (-23°C) is pre­ over the world's oceans associated with clouds; it persists for an extended period
dicted after 3 weeks. Subfreezing tem­ changes in atmospheric radiative trans­ because of the long residence time of the
peratures persist for several months. port are always small (cooling of ^ 3 K) particles at high altitudes. The size of the
Among the cases shown, even the small­ because of the great heat content and warming is due to the low heat capacity
est temperature decreases on land are rapid mixing of surface waters. Howev­ of the upper atmosphere, its small infra­
-5° to 10°C (cases 4, 11, and 12), er, variations in atmospheric zonal circu­ red emissivity, and the initially low tem­
enough to turn summer into winter. lation patterns (see below) might signifi­ peratures at high altitudes. The surface
Thus, severe climatological conse­ cantly alter ocean currents and upwell- cooling is the result of attenuation of the
quences might be expected in each of ing, as occurred on a smaller scale re­ incident solar flux by the aerosol clouds
these cases. The 100-MT city airburst cently in the Eastern Pacific (El Nino) (see Fig. 4) during the first month of the
scenario (case 14) produces a 2-month (62). The oceanic heat reservoir would simulation. The greenhouse effect no
interval of subfreezing land tempera­ also moderate the predicted continental longer occurs in our calculations because
1 week 1 month 4 months 1 year , | , i i i |
) ' i , )
100 290 Ambient
~ 13 temperature -
Cases:
1. Baseline, 5000 MT 11 / 1 0,
2. Low-yield airbursts, 5000 MT ,
9. 10,000-MT full exchange --A
10. 3,000-MT exchange 270 v Ъ^* pForeinetz ionfg
11. 3,000 MT counterforce pure water
12. 1,000-MT exchange
13. 300-MT SH Cases1.: Baseline, 5000 MT
14. 100-MT city attack 2. Low-yield airburst, 5000 MT
4. Baseline, dust only
250 9. 10,000-MT exchange
10. 3,000-MT exchange
11. 3,000-MT counterforce
12. 1,000-MT exchange
13. 300-MT SH
14. 100-MT city attack
4-40
ООП
~"0 100 200 300
Time after detonation (days)
Fig. 1 (left). Time-dependent hemispherically averaged vertical opti­
cal depths (scattering plus absorption) of nuclear dust and smoke
clouds at a wavelength of 550 nm. Optical depths ^ 0.1 are negligible,
~ 1 are significant, and > 2 imply possible major consequences.
Transmission of sunlight becomes highly nonlinear at optical depths
0.1 h
^ 1. Results are given for several of the cases in Table 1. Calculated
optical depths for the expanding El Chichon eruption cloud are shown
for comparison (53). Fig. 2 (right). Hemispherically averaged
surface temperature variations after a nuclear exchange. Results are
shown for several of the cases in Table 1. (Note the linear time scale,
unlike that in Fig. 1). Temperatures generally apply to the interior of
continental land masses. Only in cases 4 and 11 are the effects of fires
neglected.
0.01
Time after detonation (sec)
1286 SCIENCE, VOL. 222
solar energy is deposited above the effects is also obtained in case 22, where cent. These results show that light ab­
height at which infrared energy is radiat­ the tropospheric washout lifetime of sorption and heating in nuclear smoke
ed to space. smoke particles is increased from 10 to clouds remain high until the graphitic
Decreases in insolation for several nu­ 30 days near the ground. By contrast, carbon fraction of the smoke falls below
clear war scenarios are shown in Fig. 4. when the nuclear smoke is initially con­ a few percent.
The baseline case implies average hemi­ tained near the ground and dynamical One sensitivity test (case 29, not illus­
spheric solar fluxes at the ground ^ 10 and hydrological removal processes are trated) considers the optical effects in the
percent of normal values for several assumed to be unperturbed, smoke de­ Southern Hemisphere (SH) of dust and
weeks (apart from any patchiness in the pletion occurs much faster (case 25). But soot transported from the NH strato­
dust and smoke clouds). In addition to even in this case, some of the smoke still sphere. In this calculation, the smoke in
causing the temperature declines men­ diffuses to the upper troposphere and the 300-MT SH case 13 is combined with
tioned above, the attenuated insolation remains there for several months (66). half the baseline stratospheric dust and
could affect plant growth rates, and vigor In a set of optical calculations, the smoke (to approximate rapid global dis­
in the marine (63), littoral, and terrestrial imaginary refractive index of the smoke persion in the stratosphere). The initial
food chains. In the 10,000-MT "severe" was varied between 0.3 and 0.01. The optical depth is — 1 over the SH, drop­
case, average light levels are below the optical depths calculated for indices be­ ping to about 0.3 in 3 months. Predicted
minimum required for photosynthesis for tween 0.1 and 0.3 show virtually no average SH continental surface tempera­
about 40 days over much of the Northern differences (cases 1 and 27 in Fig. 5). At tures fall by 8 К within several weeks
Hemisphere. In a number of other cases, an index of 0.05, the absorption optical and remain at least 4 К below normal for
insolation may, for more than 2 months, depth (52) is reduced by only = 50 per­ nearly 8 months. The seasonal influence
fall below the compensation point at cent, and at 0.01, by — 85 percent. The should be taken into account, however.
which photosynthesis is just sufficient to overall opacity (absorption plus scatter­ For example, the worst consequences
maintain plant metabolism. Because nu­ ing), moreover, increases by — 5 per- for the NH might result from a spring or
clear clouds are likely to remain patchy
the first week or two after an exchange,
leakage of sunlight through holes in the
clouds could enhance plant growth activ­
ity above that predicted for average Fig. 3. Northern
cloud conditions; however, soon thereaf­ Hemisphere tropo­
ter the holes are likely to be sealed. sphere and strato­
sphere temperature dio ~
perturbations (in Kel­
vins; 1 К = ГС) after
Sensitivity Tests the baseline nuclear
exchange (case 1).
The hatched area in­ 100
A large number of sensitivity calcula­
dicates cooling. Am­
tions were carried out as part of this bient pressure levels
study (75). The results are summarized in millibars are also
here. Reasonable variations in the nucle­ given.
ar dust parameters in the baseline sce­ 1000
nario produce initial hemispherically av­
Time after detonation (days)
eraged dust optical depths varying from
about 0.2 to 3.0. Accordingly, nuclear
dust alone could have a major climatic 1000 —I ■ —■ 1 —■ 1 — • —:
Cases:
impact. In the baseline case, the dust 1. Baseline, 5000 MT
opacity is much greater than the total Fig. 4. Solar energy 4. Baseline, dust only
fluxes at the ground 9. 10,000-MT exchange
aerosol opacity associated with the El
over the Northern Unperturbed 4. 100-MT city attack
Chichdn and Agung eruptions (59, 64); Hemisphere in the af­ global average 6. 5,000-MT silo "severe" case"
even when the dust parameters are as­ termath of a nuclear net insolation 7. 10,000-MT "severe" case
signed their least adverse values within exchange. Results are
given for several of
the plausible range, the effects are com­
the cases in Table 1.
parable to those of a major volcanic
(Note the linear time
explosion. scale.) Solar fluxes
Figure 5 compares nuclear cloud opti­ are averaged over the
diurnal cycle and over
cal depths for several variations of the
the hemisphere. In
baseline model smoke parameters (with
cases 4 and 16 fires
dust included). In the baseline case, it is are neglected. Also
assumed that fire storms inject only a indicated are the ap­
Compensation point
small fraction (- 5 percent) of the total proximate flux levels for photosynthesis
at which photosyn­
smoke emission into the stratosphere
thesis cannot keep
(65). Thus, case 1 and case 3 (no fire­ pace with plant respi­
storms) are very similar. As an extreme ration (compensation
excursion, all the nuclear smoke is in­ point) and at which
photosynthesis ceas­
jected into the stratosphere and rapidly
es. These limits vary Limit of photosynthesis
dispersed around the globe (case 26);
^or different species.
large optical depths can then persist for a
year (Fig. 5). Prolongation of optical Time after detonation (days)
23 DECEMBER 1983 1287
summer exchange, when crops are vul­ nately, we are unable to give an accurate the NH, was assumed. Coagulation of
nerable and fire hazards are greatest. quantitative estimate of the relevant particles reduced the average opacity
The SH, in its fall or winter, might then probabilities. By their very nature, how­ after 3 months by about 40 percent.
be least sensitive to cooling and darken­ ever, the severe cases may be the most When the adhesion efficiency of the col­
ing. Nevertheless, the implications of important to consider in the deployment liding particles was also maximized, the
this scenario for the tropical regions in of nuclear weapons. average opacity after 3 months was re­
both hemispheres appear to be serious With these reservations, we present duced by — 75 percent. In the most
and worthy of further analysis. Seasonal the optical depths for some of the more likely situation, however, prompt ag­
factors can also modulate the atmospher­ severe cases in Fig. 6. Large opacities glomeration and coagulation might re­
ic response to perturbations by smoke can persist for a year, and land surface duce the average hemispheric cloud opti­
and dust, and should be considered. temperatures can fall to 230 to 240 K, cal depths by 20 to 50 percent.
A number of sensitivity tests for more about 50 К below normal. Combined
severe cases were run with exchange with low light levels (Fig. 4), these se­
yields ranging from 1000 to 10,000 MT vere scenarios raise the possibility of Other Effects
and smoke and dust parameters assigned widespread and catastrophic ecological
more adverse, but not implausible, val­ consequences. We also considered, in less detail, the
ues. The predicted effects are substan­ Two sensitivity tests were run to de­ long-term effects of radioactive fallout,
tially worse (see below). The lower prob­ termine roughly the implications for opti­ fireball-generated NO*, and pyrogenic
abilities of these severe cases must be cal properties of aerosol agglomeration toxic gases (75). The physics of radioac­
weighed against the catastrophic out­ in the early expanding clouds. (The sim­ tive fallout is well known (2, 5, 12, 27,
comes which they imply. It would be ulations already take into account con­ 67). Our calculations bear primarily on
prudent policy to assess the importance tinuous coagulation of the particles in the the widespread intermediate time scale
of these scenarios in terms of the product dispersed clouds.) Very slow dispersion accumulation of fallout due to washout
of their probabilities and the costs of of the initial stabilized dust and smoke and dry deposition of dispersed nuclear
their corresponding effects. Unfortu­ clouds, taking nearly 8 months to cover dust (68). To estimate possible exposure
1 week 1 month 4 months 1 year 1 week 1 month 4 months 1 year
100
0.01
10° 10' 106 Ю7
Time after detonation (sec) Time after detonation (sec)
Fig. 5 (left). Time-dependent vertical optical depths (absorption plus scattering at 550 nm) of nuclear clouds, in a sensitivity analysis. Optical
depths are average values for the Northern Hemisphere. All cases shown correspond to parameter variations of the baseline model (case 1) and
include dust appropriate to it: case 3, no fire storms; case 4, no fires; case 22, smoke rainout rate decreased by a factor of 3; case 25, smoke initial­
ly confined to the lowest 3 km of the atmosphere; case 26, smoke initially distributed between 13 and 19 km over the entire globe; and case 27,
smoke imaginary part of refractive index reduced from 0.3 to 0.1. For comparison, in case 4, only dust from the baseline model is considered
(fires are ignored). Fig. 6 (right). Time-dependent vertical optical depths (absorption plus scattering at 550 nm) for enhanced cases of
explosion yield or nuclear dust and smoke production. Conditions are detailed elsewhere (15). Weapon yield inventories are identical to the
nominal cases of the same total yield described in Table 1 (cases 16 and 18 are also listed there). The "severe" cases generally include a sixfold in­
crease in fine dust injection and a doubling of smoke emission. In cases 15, 17, and 18, smoke causes most of the opacity during the first 1 to 2-
months. In cases 17 and 18, dust makes a major contribution to the optical effects beyond 1 to 2 months. In case 16, fires are neglected and dust
from surface bursts produces all of the opacity.
1288 SCIENCE, VOL. 222
levels, we adopt a fission yield fraction Meteorological Perturbations sphere which suppresses convection and
of 0.5 for all weapons. For exposure to rainfall.
only the gamma emission of radioactive Horizontal variations in sunlight ab­ Despite possible heavy snowfalls, it is
dust that begins to fall out after 2 days in sorption in the atmosphere, and at the unlikely that an ice age would be trig­
the baseline scenario (5000 MT), the surface, are the fundamental drivers of gered by a nuclear war. The period of
hemispherically averaged total dose ac­ atmospheric circulation. For many of the cooling (^ 1 year) is probably too short
cumulated by humans over several cases considered in this study, sizable to overcome the considerable inertia in
months could be =* 20 rads, assuming no changes in the driving forces are implied. the earth's climate system. The oceanic
shelter from or weathering of the dust. For example, temperature contrasts heat reservoir would probably force the
Fallout during this time would be con­ greater than 10 К between NH continen­ climate toward contemporary norms in
fined largely to northern mid-latitudes; tal areas and adjacent oceans may induce the years after a war. The CO2 input
hence the dose there could be — 2 to 3 a strong monsoonal circulation, in some from nuclear fires is not significant cli-
times larger (69, 70). Considering inges- ways analogous to the wintertime pat­ matologically (7).
tion of biologically active radionuclides tern near the Indian subcontinent. Simi­
(27, 71) and occasional exposure to local­ larly, the temperature contrast between
ized fallout, the average total chronic debris-laden atmospheric regions and ad­ Interhemispheric Transport
mid-latitude dose of ionizing radiation jacent regions not yet filled by smoke
for the baseline case could be ^ 50 rads and dust will cause new circulation pat­ In earlier studies it was assumed that
of whole-body external gamma radia­ terns. significant interhemispheric transport of
tion, plus ^ 50 rads to specific body Thick clouds of nuclear dust and nuclear debris and radioactivity requires
organs from internal beta and gamma smoke can thus cause significant climatic a year or more (2). This was based on
emitters (71, 72). In a 10,000-MT ex­ perturbations, and related effects, observations of transport under ambient
change, under the same assumptions, through a variety of mechanisms: reflec­ conditions, including dispersion of de­
these mean doses would be doubled. tion of solar radiation to space and ab­ bris clouds from individual atmospheric
Such doses are roughly an order of mag­ sorption of sunlight in the upper atmo­ nuclear weapons tests. However, with
nitude larger than previous estimates, sphere, leading to overall surface cool­ dense clouds of dust and smoke pro­
which neglected intermediate time scale ing; modification of solar absorption and duced by thousands of nearly simulta­
washout and fallout of tropospheric nu­ heating patterns that drive the atmo­ neous explosions, large dynamical dis­
clear debris from low-yield (< 1-MT) spheric circulation on small scales (77) turbances would be expected in the af­
detonations. and large scales (78); introduction of termath of a nuclear war. A rough analo­
The problem of NO* produced in the excess water vapor and cloud condensa­ gy can be drawn with the evolution of
fireballs of high-yield explosions, and the tion nuclei, which affect the formation of global-scale dust storms on Mars. The
resulting depletion of stratospheric clouds and precipitation (79); and alter­ lower martian atmosphere is similar in
ozone, has been treated in a number of ation of the surface albedo by fires and density to the earth's stratosphere, and
studies (2-4, 7, 73). In our baseline case soot (80). These effects are closely cou­ the period of rotation is almost identical
a maximum hemispherically averaged pled in determining the overall response to the earth's (although the solar insola­
ozone reduction of - 30 percent is of the atmosphere to a nuclear war (81). tion is only half the terrestrial value).
found. This would be substantially small­ It is not yet possible to forecast in detail Dust storms that develop in one hemi­
er if individual warhead yields were all the changes in coupled atmospheric cir­ sphere on Mars often rapidly intensify
reduced below 1 MT. Considering the culation and radiation fields, and in and spread over the entire planet, cross­
relation between solar UV-B radiation weather and microclimates, which would ing the equator in a mean time of - 10
increases and ozone decreases (74), UV- accompany the massive dust and smoke days (75, 82, 83). The explanation appar­
B doses roughly twice normal are ex­ injections treated here. Hence specula­ ently lies in the heating of the dust aloft,
pected in the first year after a baseline tion must be limited to the most general which then dominates other heat sources
exchange (when the dust and soot had considerations. and drives the circulation. Haberle et al.
dissipated). Large UV-B effects could Water evaporation from the oceans is (82) used a two-dimensional model to
accompany exchanges involving war­ a continuing source of moisture for the simulate the evolution of martian dust
heads of greater yield (or large multi- marine boundary layer. A heavy semi­ storms and found that dust at low lati­
burst lay downs). permanent fog or haze layer might blan­ tudes, in the core of the Hadley circula­
A variety of toxic gases (pyrotoxins) ket large bodies of water. The conse­ tion, is the most important in modifying
would be generated in large quantities by quences for marine precipitation are not the winds. In a nuclear exchange, most
nuclear fires, including CO and HCN. clear, particularly if normal prevailing of the dust and smoke would be injected
According to Crutzen and Birks (7), winds are greatly modified by the per­ at middle latitudes. However, Haberle et
heavy air pollution, including elevated turbed solar driving force. Some conti­ al. (82) could not treat planetary-scale
ozone concentrations, could blanket the nental zones might be subject to continu­ waves in their calculations. Perturba­
NH for several months. We are also ous snowfall for several months (10). tions of planetary wave amplitudes may
concerned about dioxins and furans, ex­ Precipitation can lead to soot removal, be critical in the transport of nuclear war
tremely persistent and toxic compounds although this process may not be very debris between middle and low latitudes.
which are released during the combus­ efficient for nuclear clouds (77, 79). It is Significant atmospheric effects in the
tion of widely used synthetic organic likely that, on average, precipitation SH could be produced (i) through dust
chemicals (75). Hundreds of tons of rates would be generally smaller than in and smoke injection resulting from ex­
dioxins and furans could be generated the ambient atmosphere; the major re­ plosions on SH targets, (ii) through
during a nuclear exchange (76). The maining energy source available for transport of NH debris across the mete­
long-term ecological consequences of storm genesis is the latent heat from orological equator by monsoon-like
such nuclear pyrotoxins seem worthy of ocean evaporation, and the upper atmo­ winds (84), and (iii) through interhemi­
further consideration. sphere is warmer than the lower atmo­ spheric transport in the upper tropo-
23 DECEMBER 1983 1289
sphere and stratosphere, driven by solar erally nonabsorbing. Smoke particles are data base is incomplete, and because the
heating of nuclear dust and smoke extremely small (typically < 1 |xm in problem is not amenable to experimental
clouds. Photometric observations of the radius), which lengthens their atmo­ investigation. We are also unable to fore­
El Chichon volcanic eruption cloud (ori­ spheric residence time. There is also a cast the detailed nature of the changes in
gin, 14°N) by the Solar Mesosphere Ex­ high probability that nuclear explosions atmospheric dynamics and meteorology
plorer satellite show that 10 to 20 percent over cities, forests, and grasslands will implied by our nuclear war scenarios, or
of the stratospheric aerosol had been ignite widespread fires, even in attacks the effect of such changes on the mainte­
transported to the SH after - 7 weeks limited to missile silos and other strate­ nance or dispersal of the initiating dust
(85). gic military targets. and smoke clouds. Nevertheless, the
4) Smoke from urban fires may be magnitudes of the first-order effects are
more important than smoke from collat­ so large, and the implications so serious,
Discussion and Conclusions eral forest fires for at least two reasons: that we hope the scientific issues raised
(i) in a full-scale exchange, cities holding here will be vigorously and critically
The studies outlined here suggest se­ large stores of combustible materials are examined.
vere long-term climatic effects from a likely to be attacked directly; and (ii)
5000-MT nuclear exchange. Despite un­ intense fire storms could pump smoke
References and Notes
certainties in the amounts and properties into the stratosphere, where the resi­
1. J. Hampson, Nature (London) 250, 189 (1974).
of the dust and smoke produced by nu­ dence time is a year or more. 2. National Academy of Sciences, Long-Term
Worldwide Effects of Multiple Nuclear-Weapon
clear detonations, and the limitations of 5) Nuclear dust can also contribute to Detonations (Washington, D.C., 1975).
models available for analysis, the follow­ the climatic impact of a nuclear ex­ 3. R. С Whitten, W. J. Borucki, R. P. Turco,
Nature (London) 257, 38 (1975).
ing tentative conclusions may be drawn. change. The dust-climate effect is very 4. M. С MacCracken and J. S. Chang, Eds.,
1) Unlike most earlier studies [for in­ sensitive to the conduct of the war; a Lawrence Livermore Lab. Rep. UCRL-51653
(1975).
stance, (2)], we find that a global nuclear smaller effect is expected when lower 5. J. C. Mark, Annu. Rev. Nucl. Sci. 26, 51 (1976).
6. K. N. Lewis, Sci. Am. 241, 35 (July 1979).
war could have a major impact on cli­ yield weapons are deployed and air-
7. P. J. Crutzen and J. W. Birks, Ambio 11, 114
mate—manifested by significant surface bursts dominate surfai|e land bursts. (1982).
8. L. W. Alvarez, W. Alvarez, F. Asaro, H. V.
darkening over many weeks, subfreezing Multiburst phenomena might enhance Michel, Science 208, 1095 (1980); W. Alvarez,
land temperatures persisting for up to the climatic effects of nuclear dust, but F. Asaro, H. V. Michel, L. W. Alvarez, ibid.
216, 886 (1982); W. Alvarez, L. W. Alvarez, F.
several months, large perturbations in not enough data are available to assess Asaro, H. V. Michel, Geol. Soc. Am. Spec.
global circulation patterns, and dramatic this issue. Pap. 190 (1982), p. 305.
9. R. Ganapathy, Science 216, 885 (1982).
changes in local weather and precipita­ 6) Exposure to radioactive fallout 10. O. B. Toon et al., Geol. Soc. Am. Spec. Pap.
tion rates—a harsh "nuclear winter" in may be more intense and widespread 190 (1982), p. 187; J. B. Pollack, О. В. Toon, T.
P. Ackerman, C. P. McKay, R. P. Turco, Sci­
any season. Greatly accelerated inter- than predicted by empirical exposure ence 219, 287 (1983).
11. Under the sponsorship of the Defense Nuclear
hemispheric transport of nuclear debris models, which neglect intermediate fall­ Agency, the National Research Council (NRC)
in the stratosphere might also occur, out extending over many days and of the National Academy of Sciences has also
undertaken a full reassessment of the possible
although modeling studies are needed to weeks, particularly.when unprecedented climatic effects of nuclear war. The present
quantify this effect. With rapid inter- quantities of fission debris are released analysis was stimulated, in part, by earlier NRC
interest in a preliminary estimate of the climatic
hemispheric mixing, the SH could be abruptly into the troposphere by explo­ effects of nuclear dust.
subjected to large injections of nuclear sions with submegaton yields. Average 12. Office of Technology Assessment, The Effects
of Nuclear War (OTA-NS-89, Washington,
debris soon after an exchange in the NH mid-latitude whole-body gamma-ray D.C., 1979).
13. J. E. Coggle and P. J. Lindop, Ambio 11, 106
Northern Hemisphere. In the past, SH doses of up to 50 rads are possible in a (1982).
effects have been assumed to be minor. 5000-MT exchange; larger doses would 14. S. Bergstrom et al., "Effects of nuclear war on
health and health services," WHO Publ. A36.12
Although the climate disturbances are accrue within the fallout plumes of radio­ (1983).
expected to last more than a year, it active debris extending hundreds of 15. R. P. Turco, O. B. Toon, T. P. Ackerman, J. B.
Pollack, С Sagan, in preparation.
seems unlikely that a major long-term kilometers downwind of targets. These 16. R. P. Turco, P. Hamill, O. B. Toon, R. С
climatic change, such as an ice age, estimates neglect a probably significant Whitten, С S. Kiang, J. Atmos. Sci. 36, 699
(1979); NASA Tech. Pap. 1362 (1979); R. P.
would be triggered. internal radiation dose due to biological­ Turco, O. B. Toon, P. Hamill, R. С Whitten, J.
Geophys. Res. 86, 1113 (1981); R. P. Turco, O.
2) Relatively large climatic effects ly active radionuclides. B. Toon, R. С Whitten, Rev. Geophys. Space
could result even from relatively small 7) Synergisms between long-term nu­ Phys. 20, 233 (1982); R. P. Turco, О. В. Toon,
R. С Whitten, P. Hamill, R. G. Keesee, /.
nuclear exchanges (100 to 1000 MT) if clear war stresses—such as low light Geophys. Res. 88, 5299 (1983).
urban areas were heavily targeted, be­ levels, subfreezing temperatures, expo­ 17. O. B. Toon, R. P. Turco, P. Hamill, С S.
Kiang, R. С Whitten, J. Atmos. Sci. 36, 718
cause as little as 100 MT is sufficient to sure to intermediate time scale radioac­ (1979); NASA Tech. Pap. 1363 (1979).
18. O. B. Toon and T. P. Ackerman, Appl. Opt. 20,
devastate and burn several hundred of tive fallout, heavy pyrogenic air pollu­ 3657 (1981); T. P. Ackerman and О. В. Toon,
the world's major urban centers. Such a tion, and UV-B flux enhancements— ibid., p. 3661; J. N. Cuzzi, T. P. Ackerman, L.
С Helme, J. Atmos. Sci. 39, 917 (1982).
low threshold yield for massive smoke aggravated by the destruction of medical 19. Prediction of circulation anomalies and atten­
emissions, although scenario-dependent, facilities, food stores, and civil services, dant changes in regional weather patterns re­
quires an appropriately designed three-dimen­
implies that even limited nuclear ex­ could lead to many additional fatalities, sional general circulation model with at least the
changes could trigger severe aftereffects. and could place severe stresses on the following features: horizontal resolution of 10°
or better, high vertical resolution through the
It is much less likely that a 5000- to global ecosystem. An assessment of the troposphere and stratosphere, cloud and precip­
itation parameterizations that allow for excur­
10,000-MT exchange would have only possible long-term biological conse­
sions well outside present-day experience, abili­
minor effects. quences of the nuclear war effects quan­ ty to transport dust and smoke particles, an
interactive radiative transport scheme to calcu­
3) The climatic impact of sooty smoke tified in this study is made by Ehrlich et late dust and smoke effects on light fluxes and
from nuclear fires ignited by airbursts is al. (86). heating rates, allowance for changes in particle
sizes with time and for wet and dry deposition,
expected to be more important than that Our estimates of the physical and and possibly a treatment of the coupling be­
of dust raised by surface bursts (when chemical impacts of nuclear war are nec­ tween surface winds and ocean currents and
temperatures. Even if such a model were avail­
both effects occur). Smoke absorbs sun­ essarily uncertain because we have used able today, it would not be able to resolve
questions of patchiness on horizontal scales of
light efficiently, whereas soil dust is gen­ one-dimensional models, because the less than several hundred kilometers, of local -
1290 SCIENCE, VOL. 222
ized perturbations in boundary-layer dynamics, ever, in cases where the total amount of submi- noncombustible surface dust and explosion de­
or of mesoscale dispersion and removal of dust crometer volcanic material that remained in the bris (such as pulverized plaster) might be lifted
and smoke clouds. stratosphere could be determined, climate mod­ in addition to the smoke particles.
20. Advisors, Ambio 11, 94 (1982). els have been applied and tested [J. B. Pollack et 50. This assumes an average graphitic carbon mass
21. R. T. Pretty, Ed., Jane's Weapon Systems, al., J. Geophys. Res. 81, 1071 (1976)]. We used fraction of about 30 to 50 percent, for a pure
1982-1983 (Jane's, London, 1982). such a model in this study to predict the effects carbon imaginary refractive index of 0.6 to 1.0
22. The Military Balance 1982-1983 (International of specific nuclear dust injections. [J. T. Twitty and J. A. Weinman, J. Appl.
Institute for Strategic Studies, London, 1982). 35. E. Ishikawa and D. L. Swain, Translators, Hiro­ Meteorol. 10, 725 (1971); S. Chippett and W. A.
23. World Armaments and Disarmament, Stock­ shima and Nagasaki: The Physical, Medical Gray, Combust. Flame 31, 149 (1978)]. The real
holm International Peace Research Institute and Social Effects of the Atomic Bombings part of the refractive index of pure carbon is
Yearbook 1982 (Taylor & Francis, London, (Basic Books, New York, 1981). 1.75, and for many oils is 1.5 to 1.6. Smoke
1982). 36. At Hiroshima, a weapon of roughly 13 KT particles were assigned an average density of 1
24. R. Forsberg, Sci. Am. 247, 52 (November 1982). created a fire over — 13 km2. At Nagasaki, g/cm3 (С. К. McMahon, paper presented at the
25. The unprecedented difficulties involved in con­ where irregular terrain inhibited widespread fire 76th Annual Meeting, Air Pollution Control As­
trolling a limited nuclear exchange are discussed ignition, a weapon of roughly 22 KT caused a sociation, Atlanta, Ga., 19 to 24 June 1983).
by, for example, P. Bracken and M. Shubik fire over — 7 km2. These two cases suggest that Solid graphite has a density = 2.5 g/cm3, and
[Technol. Soc. 4, 155 (1982)] and by D. Ball low-yield (=s 1-MT) nuclear explosions can most oils, =s 1 g/cm3.
[Adelphi Paper 169 (International Institute for readily ignite fires over an area of — 0.3 to 1.0 51. A number of targets with military, economic, or
Strategic Studies, London, 1981)]. km2/KT—roughly the area contained within the political significance can also be identified in
26. G. Kemp, Adelphi Paper 106 (International In­ = 10 cal/cm2 and the — 2 psi overpressure con­ tropical northern latitudes and in the SH (20).
stitute for Strategic Studies, London, 1974). tours (27). 52. Attenuation of direct sunlight by dust and smoke
27. S. Glasstone and P. J. Dolan, Eds., The Effects 37. A. Broido, Bull. At. Sci. 16, 409 (1960). particles obeys the law 7//0 = ехр^т/и^), where
of Nuclear Weapons (Department of Defense, 38. С F. Miller, "Preliminary evaluation of fire т is the total extinction optical depth due to
Washington, D.C., 1977). hazards from nuclear detonations," SRI (Stan­ photon scattering and absorption by the parti­
28. The areas cited are subject to peak overpres­ ford Res. Inst.) Memo. Rep. Project IMU-4021- cles and no is the cosine of the solar zenith
sures s 10 to 20 cal/cm2. 302(1962). angle. The optical depth depends on the wave­
29. A 1-MT surface explosion ejects ~ 5 x 106tons 39. R. U. Ayers, Environmental Effects of Nuclear length of the light and the size distribution and
of debris, forming a large crater (27). Typical Weapons (HI-518-RR, Hudson Institute, New composition of the particles, and is generally
soils consist of = 5 to 25 percent by weight of York, 1965), vol. 1. calculated from Mie theory (assuming equiva­
grains =s 1 jxm in radius [G. A. D'Almeida and 40. S. B. Martin, "The role of fire in nuclear war­ lent spherical particles). The total light intensity
L. Schutz, J. Climate Appl. Meteorol. 11, 233 fare," United Research Services Rep. URS-764 at the ground consists of a direct component and
(1983); G. Rawson, private communication]. (DNA 2692F) (1974). a diffuse, or scattered, component, the latter
However, the extent of disaggregation of the soil 41. DCPA Attack Environment Manual (Depart­ usually calculated with a radiative transfer mod­
into parent grain sizes is probably ^ 10 percent ment of Defense, Washington, D.C., 1973), el. The extinction optical depth can be written as
[R. G. Pinnick, G. Fernandez, B. D. Hinds, chapter 3. т = XML, where X is the specific cross section
Appl. Opt. 11, 95 (1983)] and would depend in 42. FEMA Attack Environment Manual (CPG 2- (m2/g paniculate), M the suspended particle
part on soil moisture and compaction. 1A3, Federal Emergency Management Agency, mass concentration (g/m3), and L the path length
30. A 1-MT surface explosion vaporizes — 2 x Ю4 Washington, D.C., 1982), chapter 3. (m). It is the sum of a scattering and an absorp­
to 4 x Ю4 tons of soil (27), which is ingested by 43. H. L. Brode, "Large-scale urban fires," Pacific- tion optical depth (т = TS + та). Fine dust and
the fireball. Some silicates and other refractory Sierra Res. Corp. Note 348 (1980). smoke particles have scattering coefficients
materials later renucleate into fine glassy 44. D. A. Larson and R. D. Small, "Analysis of the A's = 3 to 5 m2/g at visible wavelengths. Howev­
spheres [M. W. Nathans, R. Thews, I. J. Rus­ large urban fire environment," Pacific Sierra er, the absorption coefficients Xa are very sensi­
sell, Adv. Chem. Ser. 93, 360 (1970)]. Res. Corp. Rep. 1210 (1982). tive to the imaginary part of the index of refrac­
31. Aqu a1n-tMitiTe s osuf rdfaucset ovexerp laons iaorne a roafi s^e s 10s0ig knmifi2c abnyt 45. tUiorbnsa ne xacnede dsuinbgu r1b0a0n, 0a0r0e a(sa boof ucti t2ie3s0 0w iwtho rpldowpuidlae­) tFioorn .s mFoork etsy,p Xica alc asno ivl arpya rftriocmle s~, J0fa. 1 st o 01.10 mm22//gg,.
"popcorning," due to thermal radiation, and by are surveyed in (15). Also discussed are global roughly in proportion to the volume fraction of
saltation, due to pressure winds and turbulence reserves of flammable substances, which are graphite in the particles. Occasionally, specific
(27). Much of the dust is sucked up by the shown to be roughly consistent with known extinction coefficients for smoke are given rela­
afterwinds behind the rising fireball. Size sorting rates of production and accumulation of com­ tive to the mass of fuel burned; then X implicitly
should favor greatest lifting for the finest parti­ bustible materials. P. J. Crutzen and I. E. Gal- includes a multiplicative emission factor (grams
cles. The quantity of dust lofted would be sensi­ bally (in preparation) reach similar conclusions of smoke generated per gram of fuel burned).
tive to soil type, moisture, compaction, vegeta­ about global stockpiles of combustibles. 53. R. P. Turco, O. B. Toon, R. С Whitten, P.
tion cover, and terrain. Probably > 1 x Ю5 tons 46. Smoke emission data for forest fires are re­ Hamill, Eos 63, 901 (1982).
of dust per megaton can be incorporated into the viewed by D. V. Sandberg, J. M. Pierovich, D. 54. J. A. Ogren, in Particulate Carbon: Atmospher­
stabilized clouds in this manner. G. Fox, and E. W. Ross ["Effects of fire on ic Life Cycle, G. T. Wolflf and R. L. Klimisch,
32. R. G. Gutmacher, G. H. Higgins, H. A. Tewes, air," U.S. Forest Serv. Tech. Rep. WO-9 Eds. (Plenum, New York; 1982), pp. 379-391.
Lawrence Liver more Lab. Rep. UCRL-14397 (1979)]. Largest emission factors occur in in­ 55. To estimate the wildfire area, we assume that 25
(1983); J. Carpenter, private communication. tense large-scale fires where smoldering and percent of the total nonurban yield, or 1000 MT,
33. M. W. Nathans, R. Thews, I. J. Russell [in (30)]. flaming exist simultaneously, and the oxygen ignites fires over an area of 500 km2/MT—
These data suggest number size distributions supply may be limited over part of the burning approximately the zone irradiated by 10 cal/
that are log-normal at small sizes (:s 3 jim) and zone. Smoke emissions from synthetic organic cm2—and that the fires do not spread outside
power law (r~e) at larger sizes. Considering data compounds would generally be larger than those this zone (39). R. E. Huschke [Rand Corp. Rep.
from a number of nuclear tests, we adopted an from forest fuels [C. P. Bankston, В. Т. Zinn, R. RM-5073-TAB (1966)] analyzed the simulta­
average log-normal mode radius of 0.25 |xm, F. Browner, E. A. Powell, Combust. Flame 41, neous flammability of wildland fuels in the Unit­
a = 2.0, and an exponent, a = 4 (15). If all 273 (1981)]. ed States, and determined that about 50 percent
particles in the stabilized clouds have radii in the 47. Sooty smoke is a complex mixture of oils, tars, of all fuels are at least moderately flammable
range 0.01 to 1000 ц.т, the adopted size distribu­ and graphitic (or elemental) carbon. Measured throughout the summer months. Because — 50
tion has - 8 percent of the total mass in parti­ benzene-soluble mass fractions of wildfire percent of the land areas of the countries likely
cles :s 1 ц,т in radius; this fraction of the smokes fall in the range — 40 to 75 percent [D. to be involved in a nuclear exchange are covered
stabilized cloud mass represents ^ 0.5 percent V. Sandberg et al. in (46)]. Most of the residue is by forest and brush, which are flammable about
of the total ejecta and sweep-up mass of a likely to be brown to black (the color of smoke 50 percent of the time, the 1000-MT ignition
surface explosion and amounts to — 25 tons per ranges from white, when large amounts of water yield follows statistically.
kiloton of yield. vapor are present, to yellow or brown, when oils 56. Most of the background smoke is injected into
34. Atmospheric dust from volcanic explosions dif­ predominate, to gray or black, when elemental the lowest 1 to 2 km of the atmosphere, where it
fers in several important respects from that carbon is a major component). has a short lifetime, and consists on the average
produced by nuclear explosions. A volcanic 48. A. Tewarson, in Flame Retardant Polymeric of ^ 10 percent graphitic carbon [R. P. Turco,
eruption represents a localized dust source, Material, M. Lewin, S. M. Atlas, E. M. Pierce, O. B. Toon, R. С Whitten, J. B. Pollack, P.
while a nuclear war would involve thousands of Eds. (Plenum, New York, 1982), vol. 3, pp. 97- Hamill, in Precipitation Scavenging, Dry Depo­
widely distributed sources. The dust mass con­ 153. In small laboratory burns of a variety of sition and Re suspension, H. R. Pruppacher, R.
centration in stabilized nuclear explosion clouds synthetic organic compounds, emissions of G. Semonin, W. G. N. Slinn, Eds. (Elsevier,
is low (s 1 g/m3), while volcanic eruption col­ "solid" materials (which remained on collection New York, 1983), p. 1337]. Thus, the average
umns are so dense they generally collapse under filters after baking at 100°C for 24 hours) ranged optical depth of ambient atmospheric soot is
their own weight [G. P. L. Walker, J. Volcanol. from — 1 to 15 percent by weight of the carbon only :£ 1 percent of the initial optical depth of
Geotherm. Res. 11, 81 (1981)]. In the dense consumed; of low-volatility liquids, = 2 to 35 the baseline nuclear war smoke pall.
volcanic clouds particle agglomeration, particu­ percent; and of high-volatility liquids, — 1 to 40 57. H. E. Landsberg and J. M. Albert, Weatherwise
larly under the influence of electrical charge, percent. Optical extinction of the smoke gener­ 11, 63 (1974).
can lead to accelerated removal by sedimenta­ ated by a large^ number of samples varied from 58. H. Stommel and E. Stommel, Sci. Am. 240, 176
tion [S. N. Carey and H. Sigurdsson, J. = 0.1 to 1.5 m2 per gram of fuel burned. (June 1979).
Geophys. Res. 87, 7061 (1982); S. Braziersal., 49. In wildfires, the particle number mode radius is 59. O. B. Toon and J. B. Pollack, Nat. Hist. 86, 8
Nature (London) 301 115 (1983)]. The size distri­ typically about 0.05 ц.т [D. V. Sandberg et al., (January 1977).
bution of volcanic ash is also fundamentally in (46)]. For burning synthetics the number 60. H. H. Lamb, Climate Present, Past and Future
different from that of nuclear dust [W. I. Rose et mode radius can be substantially greater, but a (Methuen, London, 1977), vols. 1 and 2.
al., Am. J. Sci. 280, 671 (1980)], because the reasonable average value is 0.1 (im [C. P. Bank­ 61. Notwithstanding possible alterations in the sur
origins of the particles are so different. The ston et al., in (46)]. Often, larger debris particles face albedo due to the fires and deposition of
injection efficiency of nuclear dust into the and firebrands are swept up by powerful fire soot (15,80).
stratosphere by megaton-yield explosions is winds, but they have short atmospheric resi­ 62. S. G. H. Philander, Nature (London) 302, 295
close to unity, while the injection efficiency of dence times and are not included in the present (1983); В. С Weare, Science 111, 947 (1983).
fine volcanic dust appears to be very low (15). estimates (С. К. McMahon and P. W. Ryan, 63. D. H. Milne and C. P. McKay, Geol. Soc. Am.
For these reasons and others, the observed paper presented at the 69th Annual Meeting, Air Spec. Pap. /90(1982), p. 297.
climatic effects of major historical volcanic Pollution Control Association, Portland, Ore., 64. О. В. Toon, Eos 63, 901 (1982).
eruptions cannot be used, as in (2), to calibrate 27 June to 1 July 1976). Nevertheless, because 65. The stratosphere is normally defined as the
the potential climatic effect of nuclear dust winds exceeding 100 km/hour may be generated region of constant or increasing temperature
merely by scaling energy or soil volume. How­ in large-scale fires, significant quantities of fine with increasing height lying just above the tropo-
23 DECEMBER 1983 1291
sphere. The residence time of fine particles in 72. These estimates assume normal rates and pat­ effect of soot on snow and ice fields is under
the stratosphere is considerably longer than in terns of precipitation, which control the inter­ debate (J. Birks, private communication). In
the upper troposphere because of the greater mediate time scale radioactive fallout. In severe­ general, soot or sand accelerates the melting of
stability of the stratospheric air layers and the ly perturbed cases, however, it may happen that snow and ice. Soot that settles in the oceans
absence of precipitation in the stratosphere. the initial dispersal of the airborne radioactivity would be rapidly removed by nonselective filter-
With large smoke injections, however, the ambi­ is accelerated by heating, but that intermediate feeding plankton, if these survived the initial
ent temperature profile would be substantially time scale deposition is suppressed by lack of darkness and ionizing radiation.
distorted (for instance, see Fig. 3) and a "strato­ precipitation over land. 81. In the present calculations, chemical changes in
sphere" might form in the vicinity of the smoke 73. H. Johnston, G. Whitten, J. Birks, J. Geophys. stratospheric 0 and N0 concentrations cause
3 2
cloud, increasing its residence time at all alti­ Res. 78, 6107 (1973); H. S. Johnston, ibid. 82, a small average temperature perturbation com­
tudes (75). Thus the duration of sunlight attenua­ 3119(1977). pared to that caused by nuclear dust and smoke;
tion and temperature perturbations in Figs. 1 to 74. S. A. W. Gerstl, A. Zardecki, H. L. Wiser, it seems unlikely that chemically induced climat­
6 may be considerably underestimated. Nature (London) 294, 352 (1981). ic disturbances would be a major factor in a
66. Transport of soot from the boundary layer into 75. M. P. Esposito, T. O. Tiernan, F. E. Dryden, nuclear war. Tropospheric ozone concentra­
the overlying free troposphere can occur by U.S. EPA Rep. EPA-600/280-197 (1980). tions, if tripled (7), would lead to a small green­
diurnal expansion and contraction of the bound­ 76. J. Josephson, Environ. Sci. Technol. 17, 124A house warming of the surface [W. C. Wang, Y.
ary layer, by large-scale advection, and by (1983). In burning of PCB's, for example, re­ L. Yung, A. A. Lacis, T. Mo, J. E. Hansen,
strong localized convection. lease of toxic polycyclic chlorinated organic Science 194, 685 (1976)]. This might result in
67. F. Barnaby and J. Rotblat, Ambio 11, 84 (1982). compounds can amount to 0.1 percent by more rapid surface temperature recovery. How­
68. The term "intermediate" fallout distinguishes weight. In the United States more than 300,000 ever, the tropospheric 0 increase is transient
3
the radioactivity deposited between several tons of PCB's are currently in use in electrical (~ 3 months in duration) and probably second­
days and ~ 1 month after an exchange from systems [S. Miller, Environ. Sci. Technol. 17, ary in importance to the contemporaneous
"prompt" fallout (2= 1 day) and "late" fallout 11A (1983)]. smoke and dust perturbations.
(months to years). Intermediate fallout is ex­ 77. C.-S. Chen and H. D. Orville [J. Appl. Me- 82. R. M. Haberle, С. В. Leovy, J. B. Pollack,
pected to be at least hemispheric in scale and teorol. 16, 401 (1977)] model the effects of fine Icarus 50, 322(1983).
can still deliver a significant chronic whole-body graphitic dust on cumulus-scale convection. 83. During the martian dust storm of 1971-1972, the
gamma-ray dose. It may also contribute a sub­ They show that strong convective motions can IRIS experiment on Mariner 9 observed that
stantial internal dose, for example, from 131I. be established in still air within 10 minutes after suspended particles heated the atmosphere and
The intermediate time scale gamma-ray dose the injection of a kilometer-sized cloud of sub- produced a vertical temperature gradient that
represents, in one sense, the minimum average micrometer particles of carbon black, at mixing was substantially subadiabatic [R. B. Hanel et
exposure far from targets and plumes of prompt ratios =s 50 ppb by mass. Addition of excess al., Icarus 17, 423 (1972); J. B. Pollack et al., J.
fallout. However, the geographic distribution of humidity in their model to induce rainfall results Geophys. Res. 84, 2929 (1979)].
intermediate fallout would still be highly vari­ in still stronger convection; the carbon dust is 84. V. V. Alexandrev, private communication; S.
able, and estimates of the average dose made raised higher and spread farther horizontally, H. Schneider, private communication.
with a one-dimensional model are greatly ideal­ while 2S 20 percent is scavenged by the precipi­ 85. G. E. Thomas, B. M. Jakosky, R. A. West, R.
ized. The present calculations were calibrated tation. W. M. Gray, W. M. Frank, M. L. Corrin, W. Sanders, Geophys. Res. Lett. 10, 997(1983);
against the observed prompt fallout of nuclear and C. A. Stokes [J. Appl. Meteorol. 15, 355 J. B. Pollack et aL, ibid., p. 989; В. М. Jakosky,
test explosions (15). (1976)] discuss possible mesoscale (s 100 km) private communication.
69. There is also reason to believe that the fission weather modifications due to large carbon dust 86. P. Ehrlich et al., Science 222, 1293 (1983).
yield fraction of nuclear devices may be increas­ injections. 87. H. M. Foley and M. A. Ruderman, J. Geophys.
ing as warhead yields decrease and uranium 78. C. Covey, S. Schneider, and S. Thompson (in Res. 78, 4441 (1973).
processing technology improves. If the fission preparation) report GCM simulations which in­ 88. We gratefully acknowledge helpful discussions
fraction were unity, our dose estimates would clude soot burdens similar to those in our base­ with J. Berry, H. A. Bethe, С Billings, J. Birks,
have to be doubled. We also neglect additional line case. They find major perturbations in the H. Brode, R. Cicerone, L. Colin, P. Crutzen, R.
potential sources of radioactive fallout from global circulation within a week of injection, Decker, P. J. Dolan, P. Dyal, F. J. Dyson, P.
salted "dirty" weapons and explosions over with strong indications that some of the nuclear Ehrlich, B. T. Feld, R. L. Garwin, F, Gilmore,
nuclear reactors and fuel reprocessing plants. debris at northern mid-latitudes would be trans­ L. Grinspoon, M. Grover, J. Knox, A. Kuhl, C.
70. J.. Knox (Lawrence Livermore Lab. Rep. ported upward and toward the equator. Leovy, M. MacCracken, J. Mahlman, J. Mar-
UCRL-89907, in press) reports fallout calcula­ 79. R. С Eagan, P. V. Hobbs, L. F. Radke, J. Appl. cum, P. Morrison, E. Patterson, R. Perret, G.
tions which explicitly account for horizontal Meteorol. 13, 553 (1974). Rawson, J. Rotblat, E. E. Salpeter, S. Soter, R.
spreading and transport of nuclear debris 80. С Sagan, O. B. Toon, and J. B. Pollack [Sci­ Speed, E. Teller, and R. Whitten on a variety of
clouds. For a 53QO-MT strategic exchange, ence 206, 1363 (1979)] discuss the impact of subjects related to this work. S. H. Schneider,
Knox computes average whole-body gamma-ray anthropogenic albedo changes on global climate. С Covey, and S. Thompson of the National
doses of 20 rads from 40° to 60°N, with smaller Nuclear war may cause albedo changes by burn­ Center for Atmospheric Research generously
average doses elsewhere. Hot spots of up to 200 ing large areas of forest and grassland; by gener­ shared with us preliminary GCM calculations of
rads over areas of ~ 106 km2 are also predicted ating massive quantities of soot which can settle the global weather effects implied by our smoke
for intermediate time scale fallout. These calcu­ out on plants, snowfields, and ocean surface emissions. We also thank the almost 100 partici­
lations are consistent with our estimates. waters; and by altering the pattern and extent of pants of a 5-day symposium held in Cambridge,
71. H. Lee and W. E. Strope [Stanford Res. Inst. ambient water clouds. The nuclear fires in the Mass., 22 to 26 April, for reviewing our results;
Rep. EGU2981 (1974)] studied U.S. exposure to baseline case consume an area — 7.5 x 105 km2, that symposium was organized by the Confer­
transoceanic fallout generated by several as­ or only — 0.5 percent of the global landmass; it ence on the Longterm Worldwide Biological
sumed Sino-Soviet nuclear exchanges. Taking is doubtful that an albedo variation over such a Consequences of Nuclear War under a grant
into account weathering of fallout debris, pro­ limited area is significant. All the soot in the from the W. Alton Jones Foundation. Special
tection by shelters, and a 5-day delay before baseline nuclear war case, if spread uniformly thanks go to Janet M. Tollas. for compiling
initial exposure, potential whole-body gamma- over the earth, would amount to a layer —0.5 information on world urbanization, to May Liu
ray doses :s 10 rads and internal doses ^ 10 to \im thick. Even if the soot settled out uniformly for assistance with computer programming, and
100 rads, mainly to the thyroid and intestines, on all surfaces, the first rainfall would wash it to Mary Maki for diligence in preparing the
were estimated. into soils and watersheds. The question of the manuscript.
1292 SCIENCE, VOL. 222