Click for next page ( 18


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 17
4 Dust Part of the particulate burden of the atmosphere that would result from a nuclear war would be a collection of recondensed rock and metal vapor, small spheres of glass formed from quenched rock melt, and unmelted mineral and rock particles, all loosely termed "dust. Surface bursts are much more effective than air bursts or buried explosions in raising dust to great altitude; explosions in the megaton range would loft dust into the stratosphere, where its residence time could be many months. (A general discussion of the dynamics of atmospheric nuclear explosions can be found in Glasstone's Effects of Nuclear Weapons (Glasstone and Dolan, 1977) and in Brode's (1968) review.) As will be detailed below, estimates of the dust that would be lofted by a high-~ield surface burst lie in the range 0.2 to 0.5 teragrams (1 Tg = 10 2 g ~ 106 metric tons) per megaton (Mt) of explosion energy; a likely value would be 0.3 Tg/Mt. For a nuclear exchange including 1500 Mt of surface bursts, the total lofted mass could be 330 to 825 Tg. Roughly 8 percent of the dust would be in the submicron radius range. percent to 20 percent. A plausible value for the mass of lofted submicron dust would be 40 Tg. The following sections of this chapter summarize the dynamics of nuclear clouds, the source mechanisms of dust, the available data that give estimates for dust lofting, and finally, an analysis of dust lofting in this baseline case. _ The submicron fraction could range from a few NUCLEAR CLOUD DYNAMICS Unlike soot from the long-lasting fires, dust from a nuclear explosion would be lofted to its stabilization altitude within 3 to 4 min. and, once a nuclear attack stopped, there would be no additional sources. Dust has an appreciable effect on climate only if it is of small size (submicron, or less than one micrometer (1 um) in radius) and if it is lofted to the stratosphere, where residence times are appreciable. (An altered state of the atmosphere would make estimates of residence times less certain. Consideration of dust lofted to all altitudes is required in climate simulations.) Lofting into the stratosphere requires a substantial explosion energy, a yield above roughly 1 Mt. 17

OCR for page 17
18 Most of the following discussion will be based on idealized calculations for 1-Mt surface bursts (Zinn, 1973; Horak et al., 1982; Horak and Rodis, 1983~. The characteristics of nuclear explosions have been understood since the early years of the nuclear age, but detailed unclassified accounts have appeared only in recent years. Zinn (1973) has calculated the early growth of a 1-Mt fireball, assuming a hypothetical device with a mass of 1 metric ton. The radiation temperature of the exploding bomb was chosen to be 1.6 keV (1.9 x 107 R); the explosion produced a stream of X-rays with a peak in its blackbody distribution at about 4.5 keV (5.2 x 107 K). In air at room temperature, the mean-free path for such photons is a small fraction of a meter, so the first photons emitted heated the air around the bomb, causing the opacity of this layer to drop. Subsequent photons traveled freely across the heated layer until they reached the edge, where they, in turn, were absorbed. In this way the fireball grew until the emission of X-rays by the bomb stopped; Zinn assumed this happened at 0.1 microsecond (ps). The fireball was still very hot at the end of the X-ray deposition phase, with a temperature of about 1 keV (1.2 x 107 K), and was itself a source of high-energy photons. The fir ebal1 grew by radiative (photon) transport until the growth rate was no longer large in comparison with the sound speed in the heated gas. Until this time, any hydrodynamic signal that formed at the outer pressure discontinuity would be outrun by the radiative expansion. By 50 ps, however, the fireball had cooled sufficiently that a shock wave formed. The fireball radius at that time was about 50 m. Subsequent growth was dominated by hydrodynamic phenomena, even though the fireball eventually lost about 30 to 40 percent of the total yield in the form of optical photons. By the time the fireball radius had doubled to 100 m (at about 2 milliseconds (ma)), the expansion could be described by classical blast-wave theory (Bethe et al., 1947; Taylor, 1950; Brode, 1955; Sedov, 1959~. Fireball expansion reduced the internal pressure (radiative losses contributed a relatively minor reduction) until, at a few seconds, the pressure reached atmospheric levels. The shock wave continued to expand and weaken, but fireball growth stopped at a radius of about 1 km. For a 1-Mt surface burst, fireball growth would have stopped at about 1.3 km; the fireball would behave approximately as if its yield were 2 Mt. because the ground surface acts as a mirror (see Glasstone and Dolan, 1977, p. 71~. Because the density of the heated air in a fireball is low (about 4 percent of atmospheric density), the fireball is so buoyant that it would rise a distance equal to its own radius in the first 10 s. Deformation due to buoyancy and, in the case of a low air burst, to the effects of the shock wave reflected from the surface would transform the fireball into a ring vortex, which is the cap of a mushroom cloud. As the fireball rises, air near the ground would move inward (the "afterwind.) and then up, creating the stem of a mushroom cloud. Details of fireball rise (typically at about 100 m/s) are determined by the density structure of the atmosphere, entrainment rates, wind shears, atmospheric humidity, and other factors (Sowle, 1975~. However, the general character of the rise is a result of balance

OCR for page 17
19 between the downward force of the drag interaction between the fireball and the atmosphere and the upward buoyant force as follows: 1 pav2K~R2 = 4 OR peg, 2 3 t4.1) where Pa is ambient atmospheric density, v is the rise rate, R is a drag coefficient (K ~ 2; e.g., Pick, 1958), R is a characteristic fireball radius, and g is the acceleration of gravity. The density in the fireball has been assumed to be negligible. This reduces to v = (8Rg/3R)1/2; (4.2) for a 1-Mt burst the radius is 1 km, giving a rise rate of 115 m/s, a value in good agreement with observations. Because the density of the expanding vortex would not drop with altitude as fast as the density of the atmosphere, the fireball would eventually lose buoyancy and, after a few minutes, stop rising at the so-called stabilization altitude. Figure 4.1 (from Glasstone and Dolan, 1977, p. 32) illustrates the rise of a 1-Mt air burst, and Figure 4.2 shows the measured heights of cloud tops for a series of high-yield nuclear weapons near the equator in the Pacific. These heights are somewhat less than those quoted by Foley and Ruderman (1973) and used by Turco et al. (1983~. The present data are derived from original sources (KG & G Technical Staff, 1958~. Higher latitude bursts would show less rise, because of the different atmospheric structure (Figure 4.3~. Stabilization heights for continental bursts would also be lower than the heights observed for the Pacific tests, 25 20 15 1 0 o / 0 1 2 3 TIME AFTER EXPLOSION (minutes) 4 5 6 FIGURE 4.1 Height of the top of a cloud produced by a 1-Mt low-altitude burst {after Glasstone and Dolan, 1977~.

OCR for page 17
20 40 Top O Bottom 35 _ ~Least Squares 30 _ C] 25 ~ / 20 / 15 _ ~ << Lit I I I I I I 10 0 2 4 6 - ._ ~ o 1 1 16 18 8 10 YIELD (Mt) 12 14 FIGURE 4.2 Heights of the cloud tops and bottoms at 10 min for the high-yield Pacific tests done during operation Castle. The dashed curves are least-squares fits to the data. Stabilization altitude is reached by these clouds in about 6 to 7 min. with most of the rise occurring during the first 3 min. because of the smaller water content in fireballs over continents; water vapor condensation releases energy into the cloud, increasing buoyancy. If detailed descriptions of stabilization altitudes and stratospheric penetration unexpectedly prove to be of interest in numerical simulations, parameterization based on assumed yields, burst heights, and atmospheric structure could be derived from the model of Sowle (1975~. The preceding discussion of fireball rise applies only to explosions with total energies below about 150 Mt. Although no recent explosion, man-made or natural, has approached 150 Mt. conceivable multiburst attacks on very hard targets as well as some natural explosions such as meteor impacts could exceed this energy. Theoretical analysis suggests that at yields higher than 150 Mt. the resultant ~giant" fireball would not be confined to the lower atmosphere and could rise at speeds of kilometers per second to hundreds of kilometers altitude, perhaps carrying great quantities of dust (Jones and Sandford, 1977; Jones and Rodis, 1982; Melosh, 1982; C.E. Needham, S-Cubed, Inc., Albuquerque, unpublished numerical simulations of 500-Mt explosions, 1982~. The phenomenon may be

OCR for page 17
21 40 30 I o ~ 10 /~ Am// 30 MI/USSR / / (Seitz et al. / ~ 1968) 3 Mt/Chin~ _ , - ._.._ ~ T (Telegadas Polar Tropopause o L ,,,,,l , , , ,,,,,l , , , ,,,,,l , , ,l 0.03 0.1 1 10 50 YIELD (Mt) FIGURE 4.3 Estimates of differences in cloud rise between equatorial and high-latitude bursts (after Peterson, 1970~. understood as a ballistic rise occurring because the atmosphere is unable to confine the fireball explosion. Such a situation may account for the very large mass of dust lofted by the impact proposed by Alvarez et al. (1980, 1982) to explain the Cretaceous-Tertiary boundary claystone, and it might also arise, on a smaller scale, during concentrated nuclear attacks on hardened structures and missile silos. These situations will be discussed in more detail in the section on excursions below. DUST LOFTING BY A NUCLEAR CLOUD A rising fireball can carry gas and relatively large particles to great altitude. We have seen (Figure 4.1) that a 1-Mt burst would rise about 18 km in 3 min at an average rise rate of about 100 m/s. Such a flow could loft particles of up to a few centimeters in size. An air burst fireball rises because the buoyant energy (a potential energy deficit in the earth's gravitational field) is converted into kinetic energy of the rising mass. The rate at which this conversion occurs is paVgv, where V is the fireball volume. A surface burst would incorporate vaporized rock, molten rock, and solid particles in the low-density fireball. Lifting this mass requires the expenditure of some of the buoyancy energy. The fireball dynamics would be

OCR for page 17
22 \ significantly altered (buoyancy would be eliminated} if the added mass of dust equaled the mass deficit of the fireball; i.e., M(dust) = pad. For a 1-Mt burst the radius of the low-density central region of the fireball would be about 800 m, which gives an upper limit of 2.6 Tg to the dust load that may be lifted by buoyancy. The actual loading would be about a factor of 10 less, a probable result of such factors as the limit of energy available for vaporization of rock, poor drag coupling of the large particles that dominate the mass distribution of crater ejecta, and particle-particle interactions. Analysis of particle samples obtained by aircraft from the stabilized clouds of high-yield surface bursts detonated over water and/or coral islands indicates that the clouds loft about 0.2 Tg/Mt (Gutmacher et al., 1983~. This value applies to yields of about 1 Mt or greater. There is a suggestion in the data, particularly for the lower yield test Koon (110 kt), that the lofting efficiency increases at lower yields. Roon lofted 0.5 Tg/Mt. The uncertainties in these measurements are probably rather large, but in the absence of additional information, particularly for high-yield continental surface bursts, a plausible range of lofting efficiencies is 0.2 to 0.5 Tg/Mt, with lower values favored for higher yields. A likely value for the 0.5- to 1.0-Mt surface bursts of the baseline scenario is 0.3 Tg/Mt. The lofting efficiency is not likely to approach the extreme upper limit of 2.6 Tg/Mt; at such levels the central density would approach ambient values, buoyancy would be reduced, and dynamics would be seriously perturbed. Fireball rise is observed to be consistent with a low value of dust loading. SOURCES OF DUST If a nuclear fireball is to raise significant amounts of dust to great altitude, the burst must occur very close to the ground. One measure of the ability of a fireball to raise particles is the amount of fallout observed near the explosion. This local fallout consists mostly of the largest particles, those that cannot be long supported by the flow and that fall to the ground early in the cloud rise. Glasstone and Dolan (1977) give an upper bound for a burst height giving significant local fallout at 870 ~ 4 m, where W is the yield in megatons. This is about one fireball radius. However, this does not mean that fireballs that fail to touch the ground produce no fallout and loft no small particles to stabilization altitude. It does mean that the lofted mass drops dramatically with increasing burst height. A practical limit is given by the observation that 10-kt bombs suspended at an altitude of 450 m beneath tethered balloons produced dusty stems that did not merge with the ring vortex (the mushroom cap). If this behavior scales with fireball radius and, hence, as W0 4, a 1-Mt burst would not produce a stem/cap connection for a burst height above 3 km. Because this is near the detonation altitude of maximum blast damage, many bursts in a war would produce only dusty stems that would not connect with the fireball. Note that of the mechanisms described below, only "sweep-up" and thermal dust are likely to contribute to stem dust.

OCR for page 17
23 For bursts in the air, those very close to the ground (nsurface bursts.} are most effective in raising dust. If the weapon were slightly buried, the total mass in the cloud would increase dramatically, but because much of the explosion energy is deposited in the ground and there is no radiative fireball, the cloud rise would be very modest. A surface burst can be loosely defined as one close enough to the ground that the primary interaction with the soil occurs through the agency of radiative transport instead of blast. The details will depend on the radiative characteristics of the specific weapon, but from Zinn's (1973) hypothetical 1-Mt case it can be estimated that the burst height would have to be less than a few tens of meters. X-rays would be deposited in a thin layer of rock or soil and would generate an intense shock wave in the ground. Close to ground zero, rock would be vaporized by the shock; farther out, rock would be melted; and finally, at greater distances, the rock would be displaced, creating a cloud of ejecta from the forming crater. All these processes would contribute to the dust load of the fireball. There are three additional sources of dust: the metal vapors that are the physical remains of the weapon, soil lofted in the so-called Thermal layer, n and dust swept into the stem and fireball by afterwinds. These three mechanisms are not expected to be major sources of dust for surface bursts. Recondensed vaporized material is an important source of fine particles in nuclear clouds from surface bursts. Most of the vapor is derived from rock and soil. Only a modest amount of metal is contained in a ballistic missile warhead. The relative importance of the mechanisms that produce vapor from rock and soil varies with height of burst. If the bomb were exploded at or slightly below the surface, about half or more of the energy would be delivered as a strong shock propagated into the ground. Initially, this shock would be strong enough to vaporize rock. From calculations by Butkovich (1974) for underground explosions, the amount of vaporized rock produced by a surface burst may be estimated at 0.04 Tg/Mt for a dense rock target (density of 2.6 g/cm3) and approximately 0.06 Tg/Mt for a porous dry soil or a very porous dry rock target (density of 1.4 g/cm3~. In addition to vapor, a much larger mass of melted rock would be produced by the shock. For a surface burst on a dense rock target, about 0.5 to 0.6 Tg/Mt of rock would be shock melted; up to twice as much melt would be produced from porous targets. About half of the melt would be sprayed as a conical sheet out of the expanding crater. Both sides of the sheet would then be exposed to radiation from the fireball. Because temperatures in the early fireball would exceed the vaporization temperatures typical of rock melts (0.4 eV, or about 5000 K), part of the ejected melt sheet would be vaporized. In a 1-Mt explosion the temperature of the fireball would drop below typical vaporization temperatures for rock melts after about 5 s. Local fireball temperatures adjacent to the melt sheet would drop below vaporization temperatures sooner, owing to transfer of energy to rock vapor and to increased opacity near the melt sheet. The enthalpy

OCR for page 17
24 required to vaporize silica melts is of the order of 500 calories per gram (cal/g), and, if all the energy of the fireball were transferred to the rock vapor, the entire melt sheet from a dense target would be vaporized (about 0.3 Tg/Mt). The temperature of the fireball would drop below the vaporization temperature of the melt sheet long before this could happen, however. The thin leading edge of the melt sheet, which would be exposed longest and to the highest energy radiation, probably would be entirely vaporized, but negligible vaporization would occur from the late, thick trailing part of the ejecta sheet. From rough considerations of the geometry and velocity structure of the ejecta sheet and the temperature history of the fireball, it is estimated that probably no more than about one-tenth of the melt sheet, (0.03 to 0.06 Tg/Mt) would be vaporized by radiation from the fireball. The total amount of vaporized rock (shock-vaporized plus vaporized melt) expected from a surface burst therefore is of the order of 0.07 to 0.12 Tg/Mt, depending on the porosity and compressibility of the surface material. The melt would also be the source of another class of small particles after the fireball cooled below the vaporization temperature. Divergent flow and aerodynamic disruption would break up the ejected melt sheet into droplets. Some of these droplets would remain sufficiently large that they would soon fall out of the fireball, but microscopic droplets would also be formed. The aerodynamic pressure, Pa, on the leading edge of the ejecta sheet is given approximately by Pa = 1/2 pfve, where pi is the density of the fireball and ve is the velocity of the leading edge. Initial ejection velocities are of the order of 105 cm/s, and the average density of the fireball, over time, is about 10-5 g/cm3. Hence the initial aerodynamic pressure at the leading edge is of the order of 105 dyn/cm2, sufficient to convert large droplets into a fine mist {acceleration of centimeter-size droplets is of the order of 105 cm/s2~. The physics of disruption of the ejected melt sheet is complex and is not understood in detail. It appears likely, however, that the mass of material carried up in the fireball as fine droplets or mist of shock-melted material is comparable to the mass of rock vaporized. This melt generally would be quenched to glass as the fireball cools. The total mass of small particles derived from the rock vapor and melt would be roughly 0.2 Tg/Mt, an estimate in good agreement with the observations described below. The principal remaining sources of dust are solid particles ejected from the crater or swept up by the afterwinds. The size distribution of solid particles ejected from a surface burst crater is dependent on the characteristics of the target. Even from a crater produced in massive strong rock, a small fraction of the ejecta consists of micron and submicron particles. Ejecta from laboratory-scale craters in massive basalt consist about 1 percent by mass of particles finer than 10 um; about 0.1 percent is finer than 1 um (Gault et al., 1962; Moore and Gault, 19651. A detailed analysis of fragments produced by - 10-ton conventional explosion in massive Luff showed that 6 percent by

OCR for page 17
25 mass of the particles was finer than 10 um and 3 percent was finer than 1 am (E. Shoemaker, USGS, unpublished manuscript, 1983~. Most of the fine particles could be separated from the coarser fraction only by washing and ultrasonic agitation, however. About Oe5 percent of the total mass in particulates finer than 10 um and 0.1 percent in particles finer than 1 am were easily separated by simple mechanical agitation. Most fine particles ejected from surface burst craters collide with and stick to larger fragments. As an upper bound, probably no more than about 1 percent of the total mass consisting of particles smaller than 1 um is carried to stabilization altitude in the fireball from a surface burst on a strong rock target. The mass of material excavated from a crater produced by nuclear surface burst is sensitive to small differences in height of burst and to properties of the target, but is roughly of the order of 10 Tg/Mt (cf. Cooper, 1977~. Hence the mass of the solid particles finer than 1 um that are carried to stabilization altitude has an upper bound of the order of 0.1 Tg/Mt. Ejecta from craters produced in fine particulate target material, such as fine alluvium, may be expected to yield somewhat more than 0.1 Tg/Mt of fine solids entrained in the fireball, provided that the target is dry. In ejecta from wet targets, on the other hand, the mass of fine solid particles that are separated and entrained in the fireball may be less than 0.1 Tg/Mt, regardless of whether the material is strong rock or unconsolidated particles. As height of burst is increased, delivery of energy to the shock in the ground drops rapidly. The principal sources of dust become particles condensed from vapor and particles swept up from the surface. At sufficiently low height of burst, some surface material would be completely vaporized by radiation from the early fireball and later would condense to fine particles as the fireball cooled. At greater distances, only water and other relatively volatile constituents would be vaporized by optical photons from the fireball. The gas thus produced would loft solid particles and melt droplets into the fireball. Finally, as the fireball rose, the afterwinds would scour the surface. This scouring could be an important source of dust if a dry, fine particulate soil were present at the target or if previous bursts had dried, crushed, and loosened the soil and raised precursor dust clouds. The variation in mass loading of the fireball as the height of burst is increased is illustrated schematically in Figure 4.4. In conclusion, materials directly vaporized by the nuclear explosion as well as ejects melt are the principal sources of the fine particles lofted by nuclear clouds. Because these processes are relatively insensitive to soil and rock type, data from high-yield explosions on coral islands can reasonably be used to estimate the dust lofted by continental bursts. These considerations of source mechanisms suggest that the mass of particulates lofted to stabilization altitude by surface bursts would be a few times 0.1 Tg/Mt. The available data will be examined next.

OCR for page 17
0.01 . _ to By o can CD 30.000001 lo: I Z0.0000001 0.0001 0.00001 26 1 , . Sweep-UP \ EJecta Nominal Total Dust Mass :a | ! Estimated Upper and Lower Bound for Total Dust Mass I Zero Soil Mass I Height of Burst _ >_ Bomb Mass ~ End . ~ End Fireball/ ~ End Stem/ Cratering \,`1 Earth Contact I \ I,,`/ Cloud Contact I ~ ~ ~ ~ ~ 1''1 ~ ~ ~ ~ ~ ~ I'1 ~ ~ ~ ~ '~1 ~ ~ ~ ~ ~ ~ Ill 1 0 1 00 1 ,000 1 0,000 1 00,000 SCALED HEIGHT OF BURST (ft/Mt1/3) FIGURE 4.4 Dust mass per megaton lofted into the stabilized cloud as a function of height of burst showing relative contributions of blast ejecta and sweep up for bursts on or over moderately dry desert alluvium. (Provided by J. Carpenter, R&D Associates; now with Carpenter Research Corporation.) OBSERVATIONS OF NUCLEAR DUST CLOUDS The number of primary sources for data about the particles lofted to stabilization altitude by nuclear clouds is small. Early work is reviewed by Bjornerstedt and Edvar son (19631. More recent analyses include an important series of papers by Nathans et al. (1970), Heft (1970), and Gutmacher et al. (1983~. Much of the early work on mass lofting was done with fallout samples, because radioactive fallout was a concern that demanded immediate attention. The early studies (e.g., Adams and O'Conner, 1957) elucidated the types and origins of particles in the fallout samples. Unfortunately, fallout samples are generally not representative of the particle distribution in the stabilized cloud. Even in the early days of atmospheric nuclear weapons testing, samples of bomb debris were collected from the nuclear clouds with both manned and unmanned drone aircraft. The latter method was superior because it permitted sample collection as early as 30 min to 6 h after

OCR for page 17
27 the explosions. Other programs aimed at collecting samples from domestic and non-U.S. tests probed clouds from 24 h to several days after the tests. In all these programs the emphasis was on obtaining representative samples throughout the cloud as well as minimizing the effects of fractionation (variation of specific radioactivity with particle size). Typically, the early (hours) samples contained one part in 101 or 10~i of the original radioactivity. These samples were, nonetheless, shot. n The late samples (days) contained 10 to 100 times less. Generally, two types of debris-sampling systems were used. On the early thours) missions, a single sample was collected per sortie; multiple samples per sortie were collected on late (days) missions. Both systems collected complete samples; the sampling mechanisms were designed with aerodynamic characteristics that minimized the possibility of deflection of small particles away from the samplers. The different techniques were made necessary by the high radioactivity at early times; both sampling systems used the same filter medium (IPC 1478), consisting of a cellulose bed on a cotton scrim backing. The samples were subjected to several modes of analysis, including direct observation of the particles with electron microscopes, activation analysis, spectroscopy, observations of thin sections (made from the larger particles), and sedimentation analysis. Critical for the present report is the latter type of analysis, which yields information on the particle size distribution. Unfortunately, the sedimentation analysis, using a liquid medium, has two serious drawbacks: (1) the liquid tends to disassemble aggregated particles, thus increasing the fine component of the sample, and (2) the collection and counting scheme becomes increasingly inefficient at smaller particle sizes, thus decreasing the apparent fine particle component. PARTICLE SIZE DISTRIBUTIONS The principal results concerning particle size distribution are to be found in Nathans et al. (1970~: 1. The total lofted mass is about 0.2 to 0.5 Tg/Mt. 2. For particles smaller than a few microns, the size distribution is roughly log normal with a mean radius (rm) between 0.15 and 0.35 um and a standard deviation (~) of about 2 + 0.5. 3. Above a few microns the size distribution is a power function with an exponent (a) probably bounded by 3.5 and 4.2. It should be pointed out that the Nathans et al. data involve samples from only three surface nuclear tests: Johnie Boy, 0.5 kt, Nevada; Koon, 100 kt, Bikini; and Zuni, 3500 kt, Bikini. The observed dust number and mass distribution are quite variable, and size distribution parameters have been deduced from these data only in a qualitative sense. Turco et al. (1983) define the size distribution,

OCR for page 17
28 /rO ~ min(r,rO) M(r) = 3 prmno (rm ~ + 1 J o 2 ~ 1 ln(x/r )2 1 x exp Lo ~ dx + (3 pnOrO ) x dx, lmln(rlro) where M(r) is the mass in particles smaller than radius r; ra is the transition radius separating the log normal and power function portions of the size distribution; no is a normalization constant; and p is the grain density. When normalized, equation (4.3) can be integrated to give the mass fraction of particles smaller than 1-pm radius (nsubmicron mass fractions). Because these are the only particles with significant optical cross sections and long atmospheric lifetimes, the submicron fraction is an important characteristic of the size distribution. Figure 4.5 illustrates the submicron fraction (from equation (4.3~) as a function of the mean radius (rm) of the log normal part of the distribution, for a set of plausible values of the standard deviation (~) and the exponent (a) of the power law segment. These curves illustrate the large sensitivity of the "optical" impact of the dust to the size distribution parameters. For fixed values of the width (~) and exponent (a), the submicron mass fraction increases with decreasing mean particle radius (rm). For a fixed mean radius, either increasing the exponent or decreasing the width increases the submicron mass fraction. The nuclear dust data summarized above suggest that the family of curves illustrated in Figure 4.5 defines a plausible range for the submicron fraction. The integrations of equation (4.3) were cut off at 1-cm radius, because the nuclear clouds could not lift larger particles. A nominal value for the mass fraction of particles smaller than 1-pm radius is 8 percent, with a range from a few percent to perhaps 20 percent. OPTICAL PROPERTIES OF AIRBORNE DUST The optical properties of dust raised in nuclear clouds have not been directly determined. However, much information is available for volcanic dust and for natural windblown dust. The former type of dust may represent the glassy, partly melted fraction in a nuclear cloud, and windblown dust should be similar to dust swept up by the rising fireball. The visible wavelength optical properties of volcanic dust have most recently been reviewed by Patterson et al. (1983~. They estimate that the 500-nm refractive index for the stratospheric dust from E1 Chichon's 1982 eruption was n = 1.53 - 0.001 i, based on measurements of dust collected at the ground 80 km from the volcano. The imaginary part of the refractive index was nearly independent of wavelength from

OCR for page 17
29 0.26 0.24 0.22 0.20 0.18 O 0.16 a: 0.14 cn O 0.12 c: m 0.10 0.08 0.06 0.04 0.02 \ ~ \\ \ \\\ \\ \ \\ \\ \ \, an, o= 2.0 \\ \ \ \c~ 0~=4.2 \~ \ NN ^N \ ~ 0.25 0.30 0.35 0.40 0.45 0.50 O __ 1 1 1 1 1 - 0.15 0.20 MEAN PARTICLE RADIUS (rm) FIGURE 4.5 Estimates of the fraction of dust smaller than 1-pm radius for values of the mean radius (rm), the width (~) of the log normal portion of the size distribution, and the exponent (a) describing the power law portion of the distribution. Nominal values of the parameters (a = 4, ~ = 2, and Em = 0.25) give a submicron mass fraction of 8 percent.

OCR for page 17
30 350 to 700 nm and increased slightly below 350 nm. By contrast the imaginary index for Mount St. Helens dust was three times larger. The range of real indices of refraction for various volcanic glasses is about 1.48 to 1.63 (Patterson et al., 1983) and may vary systematically with SiO2 content of the rock. Pollack et al. (1973) measured the visible and infrared optical constants of a variety of volcanic and crustal rocks. They found a real index of refraction in the visible between 1.47 and 1.57. The imaginary indices of refraction ranged from 2 x 10-5 for pure volcanic glasses such as obsidian to 1 x 10-3 for a crustal rock such as andesite. Patterson (1981) has summarized the visible optical constants of a number of crustal rocks. The imaginary index tends to decline as particle size increases above 10 Am, because the quartz fraction increases. Smaller clay particles are reddish in color, and the absorption is probably due to iron, which can vary widely in form and concentration. Generally, the imaginary index at 500 nm is between 5 x 10-3 and 1 x 10-2 and varies with wavelength (A) as strongly as V3 from 300 to 700 nm. At infrared wavelengths, all volcanic and crustal rocks show considerable wavelength dependence due to the strong silicate bands located near 10 ~m. Unfortunately, Patterson (1981) does not report data beyond 20 Am, but Pollack et al. (1973) and Toon et al. (1977) have shown that there are also strong bands just beyond 20 nm. At some infrared wavelengths the imaginary optical constants may differ by an order of magnitude between various rock types. However, near the 8- to 12-pm band centers, which are the crucial wavelengths since they lie in the atmospheric window, the measurements appear to be within a factor of 2 or 3 for different rock types and different measurement techniques. Turco et al. (1983) employed a visible refractive index of 1.5 - 0.001 i for dust, which agrees well with the values found for volcanic dust. However, this imaginary index is somewhat lower than that of typical crustal debris. Since the committee concludes that the majority of the fine dust in the stabilized cloud would be vaporized or melted, it also adopts as most appropriate the refractive index of 1.5 - 0.001 i for all visible wavelengths. Turco et al. {1983) used the optical constants of basaltic glass (Pollack et al., 1973) in the infrared. These optical constants are at the high end of the range of the imaginary index at 10 Em for a variety of crustal materials but seem best suited for glassy ejecta, so the committee also adopts them in this study. DUST LOFTED IN THE BASELINE CASE The 6500-Mt baseline case includes 400 weapons of 1 Mt or greater and 2000 smaller weapons averaging 0.5 Mt detonated as surface bursts, presumably against hard targets such as silos and buried command structures.

OCR for page 17
31 TABLE 4.1 Dust Loftinga at Mid-Latitudes in the Baseline Case 1.5 Mt 1 Mt 0.5 Mt Total Weapons 200 200 20qo 2400 Yield (Mt) 300 200 1000 1500 Cloud dust (Tg) 60-150 40-100 200-500 300-780 Stem dustb (Tg) 6-15 4-10 20-50 30-75 Total dust (Tg) 66-165 44-110 220-550 330-825 Cloud top (km) 19 17 14 - Cloud bottom (km) 9 8 7 - Tropopause (km) 12 12 12 - Stratosphere dusts (Tg) 42-105 22-55 57-143 121-303 Tropospheric dustd (Tg) 24-60 22-55 163-407 209-522 Stratospheric submicron 3-8 2-4 5-11 10-23 dust (Tg) Tropospheric submicron 2-5 2-4 13-33 17-42 dust (Tg) aDust values are shown for lofting efficiencies of 0.2 Tg/Mt (first value) and 0.5 Tg/Mt (second value). bStem dust is assumed to be 10 percent of cloud dust, based on comparison of stem and cloud volumes. This is probably a significant underestimate of the dust loading in the lower parts of the stem but more accurate in the top of the stem, which contributes to the stratospheric dust totals. No adequate data base exists to reduce this uncertainty; however, stem dust is not expected to be a major factor in the global climatic effect. CThe tropospheric dust is the sum of the stem dust and the portion of the cloud dust below the troposphere. dThe submicron mass is calculated assuming 8 percent of the dust mass lies in the submicron size range. At mid-latitudes the dust lofting is characterized by parameters shown in Table 4.1. The figures are based on the assumptions that the cloud dust is uniformly distributed from the cloud top to the bottom, that the stem contains 10 percent of the cloud dust, and that the submicron mass fraction is 8 percent. The committee recognizes that these assumptions ignore the likely variation of air and dust density with altitude. More detailed estimates could be obtained from computer simulations. However' given the uncertainties in mass lofting and source description, such detail seems unwarranted. Scenarios leading to much larger dust effects would require more detailed treatment. The ranges of lofted dust are assumed to arise only from the plausible range of 0.2 to 0.5 Tg/Mt for the lofting capabilities of the nuclear clouds. The most probable value of the lofted dust is 0.3 Tg/Mt, resulting in an estimated 15 Tg of stratospheric submicron dust. If

OCR for page 17
32 the uncertainty in the submicron dust fraction is included, the overall range of uncertainty of potential dust injections increases further. EXCURSIONS The mass of submicron dust lofted into the stratosphere in the baseline case is relatively small (10 to 24 Tg) in comparison with masses in the cases studied by Turbo et al. (19833. Contributing to this difference are the smaller weapon yields and the reduced total megatonnage in surface bursts that have been assumed in the baseline case. The committee considered excursions that might increase the role of dust in postwar climatic effects. The main, 8500 Mt. excursion adds 100 20-Mt surface bursts that might be used in attacks on superhard targets. The clouds from such bursts would reach 37 km (top) and 19 km (bottom), so that virtually all the lofted dust would reach the stratosphere. The lofted mass would be 400 to 1000 Tg (600 Tg likely), with 8 percent of the mass in the submicron fraction. The committee also considered a simultaneous attack totaling 500 Mt of surface bursts against a cluster of closely spaced hard targets. As discussed earlier, the rise of the resulting giant fireball would be qualitatively different from the rise of single-megaton buoyant fireballs. The rise rates are much greater (kilometers per second, instead of 100 m/s), so that the lofting efficiency might exceed the energy-constrained maximum of 2.6 Tg/Mt expected for buoyant fireballs. For example, the impact proposed by Alvarez et al. (1980, 1982) to explain the iridium-enriched Cretaceous-Tertiary (K-T) boundary claystone apparently lofted a total of 107 Tg (1019 g) of dust. If the 10-km diameter impactor had a velocity of 30 km/s, its kinetic energy would have been about 108 Mt. Most of this energy was deposited in the target material, but perhaps 5 percent (5 x 10 Mt) appeared as thermal energy of the vaporized projectile and target material (Jones and Kodis, 1982~. The explosive expansion of this high-pressure gas created an enormous fireball that was unconfined by the atmosphere and probably provided the energy to spread the dust worldwide. The implied lofting efficiency of the Cretaceous-Tertiary fireball is roughly 2 Tg/Mt. If this efficiency is used for the 500-Mt fireball in the postulated simultaneous attack, the mass lofted to very great altitude (perhaps 100 km; C.E. Needham, S-Cubed, Inc., Albuquerque, unpublished numerical simulations of 500-Mt explosions, 1982) would be about 1000 Tg. This value is comparable with the dust lofted by the 100 20-Mt bursts in the 8500-Mt excursion. SUMbfARY The mass of submicron dust lofted into the stratosphere during a nuclear war would depend most critically on the following factors: (1) the number and individual yields of weapons used in surface bursts, (2) the lofting efficiency of the fireballs, and (3) the size distribution of particles in the stabilized cloud. Available data confirm that the

OCR for page 17
33 size distribution is roughly log normal below a few microns (rm = 0.25 Am, ~ = 2) and follow a power law (a = 4) at larger sizes. The lofting efficiency is probably 0.3 Tg/Mt (within an observed range of 0.2 to 0.5) for yields capable of reaching the stratosphere. These estimates agree well with those of Turco et al. (1983) for size distribution and lofting efficiency. For the present 6500-Mt baseline case the total mass of lofted dust would be of the order of 500 Tg, with a stratospheric submicron total of 15 Tg. Within the baseline case the latter figure is unlikely to exceed 30 Tg. An excursion involving higher yield weapons or concentrated attacks on hard targets might increase the masses to as much as 1000 Tg (total) and 80 Tg (stratospheric submicron particles). REFERENCES Adams, C.E., and J.D. O'Connor (1957) The Nature of Individual Radioactive Particles. VI. Fallout Particles from a Tower Shot, Operation Redwing. Report NRDL-TR-208. San Francisco, Calif.: U.S. Naval Radiological Defense Laboratory. Alvarez, L.W., W. Alvarez, F. Asaro, and H.W. Michael (1980) Extraterrestrial cause for the Cretaceous-Tertiary extinction. Science 208:1095-1108. Alvarez, W., L.W. Alvarez, F. Asaro, and H.W. Michael (1982) Current status of the impact theory for the terminal Cretaceous extinction. Geol. Soc. Am. Spec. Pap. 190:305. Bethe, H.A., K. Fuchs, J.O. Hirschfelder, J.L. McGee, R.E. Peierls, and J. von Neumann (1947) Blast Wave. Report LA-2000. Los Alamos, N.Mex.: Los Alamos Scientific Laboratory. Bjornerstedt, R., and K. Edvarson (1963) Physics, chemistry, and meteorology of fallout. Annul Rev. NUC1. Sci. 13:505. Brode, H.L. (1955) Numerical solutions of spherical blast waves. J. Appl. Phys. 26:766. Brode, H.L. (1968) Review of nuclear weapons effects. Annul Rev. Nucl. Sci. 18:153. Butkovich, T.R. (1974) Rock Melt from an Underground Nuclear Explosion. Report UCRL-51554. Livermore: University of California, Lawrence Livermore National Laboratory. 10 pp. Cooper, H.F. (1977) A summary of explosion cratering phenomena relevant to meteor impact events. Pages 11-44 ~n Impact and EXP10S ion Cratering, edited by D.J. Roddy, R.O. Pepin, and R.B. Merrill. New York: Pergamon Press. EG & G Technical Staff (1958) Operation Castle: Cloud Photography. Report WT-933. Boston, Mass.: U.S. Atomic Energy Commission. Foley, H.M., and M.A. Ruderman (1973) Stratospheric NO production from post-nuclear explosions. J. Geophys. Res. 78:4441. Gault, D.E., E.M. Shoemaker, and H.J. Moore (1962) Spray ejected from the lunar surface by meteoroid impact. NASA Tech. Note D-1767. 39 PPe

OCR for page 17
34 Glasstone, S., and P.J. Dolan (1977) The Effects of Nuclear Weapons (3rd edition). Washington, D.C.: U.S. Government Printing Office. Gutmacher, R.G., G.H. Higgins, and H.A. Tewes (1983) Total Mass and Concentration of Particles in Dust Clouds. Report UCRL-14397 Revision 2. Livermore: University of California, Lawrence Livermore National Laboratory. Heft, R.E. (1970) The characterization of radioactive particles from nuclear weapons tests. Radionuclides in the Atmosphere. Adv. Chem. Ser. 93:254. Horak, H.G., and J.W. Rodis (1983) RADFLO--A User's Manual. Report LA-9245. Los Alamos, N.Mex.: Los Alamos National Laboratory. Horak, H.G., E.M. Jones, M.T. Sandford II, R.W. Whitaker, R.C. Anderson, and J.W. Kodis (1982) Two-Dimensional Radiation- Hydrodynamic Calculation of a Nominal 1-Mt Nuclear Explosion Near the Ground. Report LA-9137. Los Alamos, N.Mex.: Los Alamos National Laboratory. Jones, E.M., and J.W. Kodis (1982) Atmospheric effects of large-body impacts: The first few minutes. Geol. Soc. Am. Spec. Pap. 190:175. Jones, E.M., and M.T. Sandford (1977) Numerical simulations of a very large explosion at the earth's surface with possible applications to tektites. Page 1009 ~n Impact and Explosion Cratering. New York: Pergamon Press. Melosh, H.J. (1982) The mechanics of large meteoroid impacts in the earth's oceans. Geol. Soc. Am. Spec. Pap. 190:121. Moore, H.J., and D.E. Gault (1965) The fragmentation of spheres by projectile impact. Pages 127-164 ~n Astrogeologic Studies Annual Report, July 1, 1964 to July 1, 1965, Part B: Crater Investigations. Flagstaff, Ariz.: Department of the Interior, U.S Geological Survey. Nathans, M.W. (1970) The specific activity of nuclear debris from ground surface bursts as a function of particle size. Radionuclides in the Atmosphere. Adv. Chem. Ser. 93:352. Nathans, M.W., and R. Thews (1970) The particle size distribution of nuclear cloud samples. Radionuclides in the Atmosphere. Adv. Chem. Ser. 352:360. Nathans, M.W., R. Thews, W.D. Holland, and P.A. Benson (1970) Particle size distribution in clouds from nuclear airburst. J. Geophys. Res. 75:7559. National Research Council (1975) Long-Term Worldwide Effects of Multiple Nuclear Weapons Detonations. Washington, D.C.: National Academy of Sciences. ~pik, E.J. (1958) Physics of Meteor Flight in the Atmosphere. Interscience Tracts of Physics and Astronomy 6. New York: Inter science. Patterson, E.M. (1981) Optical properties of the crustal aerosol in relation to chemical and physical characteristics. J. Geophys. Res. 86:3236-3246. Patterson, E.M., C.O. Pollard, and I. Galindo (1983) Optical properties of the ash from E1 Chichon volcano. Geophys. Res. Lett. 10:317-320.

OCR for page 17
35 Peterson, K.R. (1970) An empirical model for estimating world-wide deposition from atmospheric nuclear detonations. Health Phys. 18:357-378. Pollack, J.B., O.B. Toon, and B.N. Khare (1973) Optical properties of some terrestrial rocks and glasses. Icarus 19:372-389. Sedov, L.I. (1959) Similarity and Dimensional Methods in Mechanics. New York: Academic Press. Sowle, D.H. (1975) Implications of Vortex Theory for Fireball Motion. Report DNA-3581F. Washington, D.C.: Defense Nuclear Agency. Taylor, G.I. (1950) The formation of a blast wave by a very intense explosion. I. Theoretical discussion. Proc. Roy. Soc. Ser. A 2:159. Toon, O.B., J.B. Pollack, and C. Sagan (1977) Physical properties of the particles composing the Martian dust storm of 1971-1972. Icarus 30:663-696. Turco, R.P., O.B. Toon, T.P. Ackerman, J.B. Pollack, and C. Sagan (1983) Global Atmospheric Consequences of Nuclear War. Interim Report. Marina del Rey, Calif.: R&D Associates. 144 pp. Zinn, J. (1973) A finite difference scheme for time-dependent, spherical radiation hydrodynamics problems. J. Comp. Phys. 13:569.