UCRL- 89770 PREPRINT UCRL--8977 0 NUCLEAR WAR PRELIMINARY ESTIMATES OF TIIE CLIMATIC EFFECTS OF A NUCLEAR EXCHANGE Michael C MacCracken This paper was presented at the Third International Conference on Nuclear War Erice Sicily August 19-23 1983 October 1983 This is a reprint of a paper Intended fo1r publl•ation In a Journal or proc eedlng Sin•• •hinges may be made before publltatlo11 this preprint Is mode 1v• ll1ble with the un- derstanding that It will not be •lted or rei1ro ced without the permission of the author II t6N F - 6 3 oFl I 3 cf DE84 0027 29 _ DISCLAIMER This repon was prepared as an acc ount of work s n Government Neither the United States G po sored by •n agency of the Untied St ates employees makes any warranty express o emui nt nor any agency thereof nor any of their bility for the accuracy complet n- or imf P • or ass mes any legal liability or rcsponsi ' - • 0 r Use U1nC3S or any tnform f process dtsclosed or represents that its Id ion apparatus product or enc herein to any specific commercill l P' u u not infnnge _Prrvately owned rights Refermanufocturer or otherwise does _ _ P or service by trade name trademarit 001 • '·•·•·••••rnY corut1tute or imply 1·1s •ndoraomcnt recomme nda lion or favonng by the United St I Go and opinions of authers e•p-··-' here· a esd vcrnment or any agency thereof The views • 1n o not neccssaril Umted States Government or any gcncy thereof y state or relkct those of the NUCLEAR WAR PRELIMINARY ESTIMATES OF THE CLIMATIC EFFECTS OF A NUCLEAR EXCHANGE Michael c MacCracken Atmospheric and Geophysical Sciences Division Lawrence Livermore National Laboratory Livermore CA 94550 ABSTRACT The smoke rising from burning cities industri l areas and forests if such areas are attacked as part of a major nuclear exchange is projected to increase the hemispheric average atmospheric burden of highly abs cbent carbonaceous material by 100 to 1000 times As the smoke spreads from these fires it would prevent sunlight from reaching the surface leading to a sharp cooling of land areas over a several day period Within a few weeks the thick smoke would spread so as to largely cover the mid-latitudes of the Northern Hemisphere cooling mid-continental smoke-covered areas by perhaps a few tens of degrees Celsius Cooling of near coastal areas would be substantially less since oceanic heat capacity would help to buffer temperature changes in such regions The solar radiation not being absorbed at the surface would be absorbed by the smoke in the middle troposphere up to heights of perhaps 10 m As the smoky layer warms this heating of the upper troposphere wc• 1ld induce further mixing of the smoke up into the atmosphere where the smoke could remain even longer than the 10-20 days that normal scavengin 6 now allows The strong atmospheric stability created by the strongly wa i med smoke layer overlying the cooled surface and lower troposphere would tend to reduce precipitation over both the ocean and the land areas where evaporation would also be reduced due to aurface cooling The precipitation that does occur would likely be shallow and relatively ineffective in scavenging the higher smoke layer Thus solar absorption by the smoke and the reduction in scavenging would allow the smoke particles to remain in the 2 atmosphere for longer times thereby p obably prolonging the darkness and continental cooling for perhaps several months The net effect of a summertime nuclear exchange would be that summer conditions in mid-latitudes would turn o dark near winter-like conditions while a wintertime nuclear exchange would lead to somewhat more severe winter conditions Lower latitude temperatures would ecome more lik those in middle or higher latitudes The impacts of these climatic perturbations on society and agriculture remain to be evaluated ii Injections of dust and nitrogen dioxide into the stratosphere as a result of a hypothetical nuclear exchange involving large numbers of nuclear explosions having yields greater than about one-half megaton are projected to have an effect on the stratospheric radiation balance about equivalent to that of very large volcanic eruptions e g Krakatoa in 1883 or Tambora in 1816 The climatic cooling of the Northern Hemisphere from such injections could be one to a ew degrees Celsius with variable effects regionally resulting from possible shifts in storm tracks The calculations on which theae findings are based contain many ssumptions shortcomings and uncertainties that affect many aspects of the estimated response It seems nonetheless quite possible that if a nuclear exchange involves attacks on a very iarge number of cities and industriai areas thereby starting fires that generate as much smoke as is suggested by recent studies substantial cooling could be expected that would last weeka to months over most continental regions of the Northern Hemisphere but which may have relatively little direct effect on •' e Southern Hemisphere INTRODUCTION Crutzen and Birks 1982 and Turco et al 1983 have sug··· gested that a major nuclear exchange in which urban and forest areas are primary targets would lead to extensive fires and massive injections of smoke into the middle troposphere and above These emissions would be in addition to the nitrogen oxide and dust injections j_nto the atmosphere occurring coincident with the nuclear explosiohs A comparison of the mass of material projected to b injected as a result of a major nuclear exchange with that estimated to have resulted from major volcanic eruptions of the last two hundred years is given in Table 1 Estimates of current background loading are ala included in the table There a1·e substantial uncertainties in the estimates of emissions from both the volcanic eruptions and the nuclear exchange values for which are highly dependent on the particular exchange scensrio It appears that the expected stratospheric injection of sub-micron size nuclear-induced dust however would ' i 3 Table 1 Mass of Injected Material from a Major Nuclear Exchange Compared to Volcanic Injections and Background 10 12 g Species for Various Events Aerosol including dust Nuclear Exchange dust Tambora 1815 Krakatoa 1883 Agung 1963 Mt St Helens 1981 El Chichon 1982 Background Northern Hemisphere Nitrogen Oxides aa N Nu lear Explosions Generated Injection bomb-created Fire-Generated Injection Background Northern Hemisphere Ozone as 03 Nuc I ear Exchange Induced Change Background Soot Injection by Nuclear-Exchange Induced Fires Ci ties Forests Annual Forest Fire Injection in U S Background _j 118 hundreds so 10-20 'I 0 10-20 'I 'I 1 Troposphere sooot 200000 20000 4-8000 2000 2-4000 5-10 7 0 'I G '-600 40-1oot ' 1700 '200 0 0 150 57 ' 0 a few 'I 0 • 1-1 t The climatic effect of N 3nd dust injections iuto the troposphere and the resulting effect on tropospheric ozone are not treated in this calculation -- 1 Stratosphere -· Values are drawn from estimates given by a variety of sources including Turco et al 1983 Crutzen 1983 Lamb i970 Ellsaesser 1975 Turco et al 1982 Penner 1983 and our own evaluations For the nuclear exchange and major volcanoes tropospheric aerosol lofted is total amount rather than just amount remaining in the atmosphere 4 be comparable to injections from major volcanic eruptions that are bel eved to have led to noticeable i l°C coolings of surface temperatures over a few year period following the eruption The nuclear-generated nitrogen oxide injEct on which leads to formation of N ½ an absorber of solar radiation and destruction of ozone an absorber of solar radiation and active infrared gas also is clearly a major perturbation of background levels Of most concern however is the very large nearly inE tan·· taneous injection of soot urban fires are generally assumed to burn to near completion in a few days and forest fires in about a week which is a strong absorber of solar radiation We assume that the soot has a log-normal particle size distribution with a mode radius f 0 06 µm and a standard deviation of 2 2 Other radiative characteristics are taken from Turco et al 1983 Radiative calculations e g Crutzen and Birks 1982 Turco et al 1983 Cess 1983 suggest that the hundred to thousand fold increase in atmospheric soot concentration would prevent sunlight from reaching the surface over those latitudes whereever substantial amounto of the smoke spread These calculations also assume that the particles do not increase in size with time and will therefore not affect the infrared radiation balance an assumption that allows for the sharp cooling that is found While there are many devastating social and economic impacts of a nuclear exchange comparison of injections projected to occur from plausible but admittedly uncertain scenarios for a nu lear exchange indicate that there is the potential fo substantial impacts on the hemispheric and per'naps global radiation balance and climate Better understanding of the potential climatic effects is essential both so that there is an accurate perception of the post-exchange environment and to indicate that many nations whether elligerent or bystander will be affected significantly if a nuclear exchange does occur BACKGROUND The equitable climate of the Earth results from the complex interplay of many energy reservoirs and interactive processes that store and redistribute the incoming solar energy The composition of the atmosphere and character of the surface d termine both how much solar energy ie absorbed and is reflected to space and the flux of infrared radiation between the surface the atmosphere and space Atmospheric motions move the absorbed energy both vertically and horizontally in order to relieve dynamic and thermal instabilities through these motions cooling low latitudes and warming high latitudes Solar absorption at the surface evaporates water primarily from the ocean providing water vapor to the atmosphere that educes infrared emission to space and increases such emission to the surface Condensation of the 1- 5 water vapor in the atmosphere far from where it was evaporated both balances the net radiative cooling of the atmosphere and cleanses it of particulate matter The rapid increas of evaporation rate with temperature tends to prevent overheating of the ocean whereas the ocean's large heat capacity tends to prevent rapid cooling The3e many processes interact to form a climate that has been remarkably stable on time scales of decades annual average hemispheric temperatures have varied by only about 0 5°C over the last hundred years but which has been rather unstable on time scales of millenia e g ice sheets a few kilometers thick covered Canada only 18000 years ago when global average ocean surface temperatures were only about 2°C colder than at present The apparent cause of r limate variations on the longer scale have been relatively small latitudinal and seasonal shifts i e several percent in the solar radiation striking the top of the atmospher resulting from slight eccentricities in the orbital parameters of the Earth the relatively large climatic consequences that result occur because of augmentation of the initial causal factors by numerous feedback processes There is also some evidence that perturbations of the radiation balance resulting from volcanic injections into the st atosphere have been the cause of at least some of the climatic variability over the last hundred years e g Hansen et al 1981 Thus the suggestion th at the material injected into the atn osphere from the blasts and fires created by a major nuclear exchange will significantly perturb the solar radiation balance of the atmosphere deserves careful study METHODOLOGY 1 I J I 1 Although qualit tive analysis of the climate system can provide important insights quantita tive estimates of possible perturbations to the delicately balanc d climate system require the use f numerical models that represent at least those aspects of the physics of the climate system that become involved in the response to exchange-generated injections Given the magnitude and complexity of the DOtential perturbation such a model must treat space scales from a few to thousands of kilometers and from hours to months radiative dispersive and convective processes acting on and affected by the injected aerosol and land and ocean surface interactions There are not at present however verified models fonnulated to deal with all of these processes and scales simultaneously In the interest of estimating rather than fully assessing potential effects analyses done to date have resorted to use of available models that involve important simplifications and approximations 6 To develop an initial estimate of the effects of the soot on the radiative fluxes and the hemispheric average temperatures a one-riimensional model has been used to calculate perturbations as a function of altitude Turco et al 1983 used an extension of Manabe and Wetherald's 1967 radiative-convective model to study the time-dependent change of radiative fluxes and atmospheric temperatures assuming there was no land surface heat capacity and that the smoke aerosol was instantaneously spread over the entire Northern Hemisphere Our preliminary results described in the next section are based on a very similar one-dimensional model derived instead from our two-dimensional model MacCracken et al 1981 but allowing for a surface heat capacity These models explicitly calculate the vertical distribution of solar and infrared radiative fluxes including the effects of prescribed cloudiness a simplification in our present implementation of the perturbation to the solar radiation however will underestimate the effect somewhat at low soot concentrations and therefore may indicate a lightly faster recovery from the cooling than will actually be the case One-dimensional models can also only represent vertical heat transport with a simple lapse rate limitation and do not consider horizontal transport the hydrologic cycle or cloud formation processes The removal rates for the aerosol must also be prescribed based for example on removal rates in the unperturbed atmosphere The inability to treat the horizontal distribution of land with its low or zero surface heat capacity and ocean with its relatively high surface heat capacity resulting from ocean mixing processes is an extremely important simplification since it is likely that air flow from one region to the other will be accelerated by the greatly increased temperature differences between the two surface types and thereby moderate temperature changes over land The assumption that the aerosol is instantaneously wellmixed over the Northern Hemisphere is also important since it in effect exposes more solar radiation to absorption and scattering by the aerosol than would be the case following a nuclear exchange when thicker smoke exists over more limited regions For this reason the one-dimensional approach will overestimate the hemispheric average radiative and climatic impacts To address some of these issues we have applied higher dimensional models To improve calculation of the dispersal of the smoke we have developed the GRANTOUR trace species transport model Walton and MacCracken 1983 that uses transport and precipitation data from a 3-D general circulation model Gates and Schlesinger 1977 to spread the smoke around the globe from source regions and to scavenge the smoke by rain and snow While this model appears to give reasonable estimates of a rosol lifetimes and its results for radionuclide lifetimes compare 7 acceptably with observations refer to Knox 1983 there are a number of important simplifications and assumptions Of most importance is the assumption that the presence of the aerosol will not affect the dispersal and scavenging characteristics of the atmosphere which given the results to be shown will almost certainly overestimate removal rates Our present implementation of GRANTOUR also neglects vertical wind shear in the atmosphere which will somewhat underestimate spread of the smoke and may overestimate the unevenness or patchiness of the smoke GRANTOUR will however provide approximate estimates of the time for the smoke to spread around the mid-latitudes and to higher and lower latitudes These estimates have been used to specify the latitudinal and vertical distribution of the smoke and dust and NOx injected into the stratosphere for introduction into our two-dimensional latitude and vertical climate model MacCracken et al 1981 referred to hereafter as the LSDM This model provides 10° latitude resolution from pole to pole with nine layers in the vertical extending to about 35 km altitude Temperature water vapor surface pressure and wind fields are predicted based on approximations to the conservation equations for energy water vapor mass and momeGtum Cloud cover is calculated diagnostically as a function of relative humidity The land-ocean distribution at each latitude is represented approximately based on averaging over the appropriate fractions of several surface types each treated separately in terms of their own energy balance radiative fluxes hydrology and surfa ce temperature While this approximation will provide a better estimate of latitJdinal and hemispheric verage temperature and hydrologic cycle perturbations than a one dimensional model the regional distd bution and central continental perturbation extremes will not be properly represented hleksandrov 1983 goes a step further than our 2-D approxim tion in presenting results from a 3-D ocean-atmosphere general eirculation model the atmospheric portion of which is a derivative of the Oregon State University model used to drive GRANTOUR that inciudes proper global geography but his approach is more limited in terms of the detail of its radiative approximation and does not yet treat the time-dependent spread of the smoke Thus Aleksandrov's and our results can be viewed as complementary but recognizing that both approache1J still contain many shortcomings SPREADING OF THE SMOKE In order to allow comparison with the climate impacts estimated by Turco et al 1983 we have assU111ed that 150 Tg of soot aerosol are injected into the troposphere above the surface boundary layer from urban fires within one day of the nuclear exchange and an additional 57 Tg of soot are injected from forest 8 fires over the seven days following the exchange If spr ad evenly over t e Northern Hemisphere this amount of soot would with the aerosol optical properties that we assume have a vertical optical depth of more than four about equally divided between absorption and scattering For reference Turco et al 1983 suggest that heavy overcast conditions have an absorption optical depth of about l to 1 5 We however arbitrarily assume that the particles are injected in equal amount from four likely target regions centered on the Ohio River basin the southwestern United States the Rhine Valley and Moscow Following emission the particles are spread by the winds and scavenged primarily by precipitation For the case shown we will be using as the meteorological conditions results from a GCM simulation for January a period when zonal winds cyclogenesis and precipitation are all relatively high compared to other seasons we do not believe however that the general characteristics of the results are strongly sensitive to this choice of meteorology After three days the smoke covers the North Atlantic Ocean and most of the mid-latitude continental regions of the Northern Hemisphere see Figure 1 but most of the Northern Hemisphere sees no smoke at all even though the hemispheric average optical depth is 3 8 There are large regions where the optical depth is greater than 20 with large horizontal gradients present that might well perturb dynamical and scavenging processes The patchy nature of the smoke distr ibution is shown in Figure 2 which indicates the fractional area of the hemisphere having optical depth g eater than a given amount at various times following the exchange By day 30 normal scavenging processes would reduce the optical depth to about 1 in agreement with the rate of decrease considered by Turco et al 1983 The hemispheric distribution see Figure 3 still shows extensive patchiness with a number of small regions having optical depth greater than 5 and a rather large area with an optical depths greater than l While GRANTOUR will underestimate dispersion and spreading unless the intensity of the large scale planetary waves is greatly diminished the predicted heterogeneous character of the distribution oattern is likely to be generally correct indicating that particular locations may experience large day-to-day variations in smoke intensity By day 60 of the simulation see Figure 4 only a few smoky patches remain that have -r 1 most occurring at high latitudes because at lower latitudes precipitation processes have been able to more effectively cleanse the atmosphere The hemispheric average optical depth is however about 0 5 so after 60 days the smoke is more uniformly spread than at earlier timeR see Figure 2 Because initial studies of the effects of the soot on the radiation balance indicate that the upper tropos1 here may warm g Fig 1 Hemispheric distributi n of smoke-induced optical depth 3 days after the hypothetical nuclear exchange forest fires are assumed to be still burning assuming precipitation '• avenging at non-perturbed wintertime rates As assumed by Turco et al 1983 the hemispheric average optical depth is 3 8 but the pattern is not hemispherically homogeneous thereby stabilizing the atmosphere and leading to reduced pr cipitation scavenging of the aerosol these initial GRANTOUR calculations probably unde estimate particle lifetimes Although it would be desirable to investigate the dependence of particle concentrations on scavenging rates with a coupled model this has not yet been possible Therefore to investigate the effect of reduced scavei ging we have carried out a simulation in which the precipitation rate was arbitrarily multiplied by e-T 3 where Tis the local optical depth This dependence is actually somewhat less than indicated by a comparison of observed solar radiation absorbed at the surface and observed precipitation rate see Figure 5 an analog that should be viewed with caution but which does generai y exhibit the postulated relationship Figure 6 shows the hemispheric optical depth distr bution 30 days after the hypothetical e hange assuming damped scavenging 10 r 45 I a '· 30 25 3 cays 20 JS 30days 60doys -- ---c -r--- - - --- 0 L 0 1 0 2 0 4 Fraction Fig 2 o s 0 6 or hemisphere 0 7 0 8 0 9 1 0 Fractional area of the Northern Hemisphere erperiencing optical depths greater than the value indicated at various times following a nuclear exchange assuming continued precipitation scavenging t non-perturbed wintertime rates and wintertime circulation patterns Compared to the results shown in Figure 3 optical depth remains significantly greater over a large fraction of the hemisphere Similarly at day 6v as shown by contrasting Figure 7 to Figure 4 the optical depth remains significantly higher when scMvenging is reduced Note that the aerosol becomes better spreacl out in these cases contrast Figure 8 and Figure 2 but that patchiness is still strongly evident A comparison Qf these two cases anJ of a third case in which precipitation scavenging was reduced bye-• are shown in Figure 9 Quite clearly if stabilization and a reduction in scavenging rates occur as a result of solar absorption by the soot particles particle lifetime will increas greatly ani prolong any climatic perturbation ONE-DIMENSIONAL ESTIMATES OF TEMPERATURE CHANGE We have used a one-dimensional radiative convectiv model with an interactive surface to develop an initial stimate of the clim atic effects of injections of soot dust and nitroger 1 dioxide and of the resulting reduction in stratospheric ozone 11 J 0 0 _ ' Fig 3 Hemispheric distribution of smoke-induced optical depth on day 30 following the hypothetical nuclear exchange assuming precipitation scavenging at non-perturbed wintertime rates Hemispheric average opti al depth is 1 1 but the pattern remains uneven which is a result of scavenging in individual storm systems Smoke concentration as a function of time is based on the hemispheric integral of the GRANTOUR results A dust injection of 118 Tg and a nitrogen dioxide injection of 8 3 Tg N are also assumed to be spread instantaneously over the Northern Hemisphere stratosphere with scavenging occurring only very sloYly since the stratospheric residence time is assumed to range from 4 1 2 to 8 months depending on latitude To account for ozone reduction by nitrogen oxides the ozone concentration is assumed to be multiplied by 0 3 O r·-a NO z I where the coefficient a is based on results from the LLNL one-dimensions l chemical kinetics model Luther 1983 Assuming normal wintertime scavenging rates for the soot the change in land surface air temperature as a function of time after the exchange is shown in Figure 10 Because we are using a one-dimensional model these res lts can only loosely be interpreted in terms of a change in the annual average hemispheric 12 r I II I I Fig 4 Hemispheric distribution of smoke-induced optical depth on day 60 following the hypothetical nuclear exchange assuming precipitation scavenging at non-perturbed wintertime rates Hemispheric average optical depth is 0 5 with relatively few areas with values greater than 1 indicating greater uniformity of soot concentrations than at earlier times land surface temperature as a function of time Tne temperature is seen to drop rapidly decreasing more than 30° C within two weeks aE1suming constant rather than patchy smoke cover This cooling time is controlled by the heat capac ity of the lower atmosphere which is low This rest lt compares quite well with the results of Turco et al 1983 The relative warming that occurs from two to six weeks after the exchange is primarily a result of the scavenging of the soot from the atmosphere a process with a time constant of 2 to 3 weeks The slower recovery of temperature beyond six weeks reflects the effect of the stratospheric injections which are scavenged from the atmosphere more slowly than the tropospheric soot To test the robustness of this predicted perturbation we have conducted a variety of sensitivity experiments with the onedimensional model as listed in Table 2 The second and third 13 I 3 200 0 uo 15 - t Subtropics •• I 'O C bD 100 0 rJ J 0 I 0 • • - 5i 'O f •✓ ✓• •• ✓ 2 3 4 5 Precipitation rate mm day Fig 5 Relationship uf solar radiation absorbed at the surface in our zonal climate model to observed precipitation rate Jaeger 1976 as a function of latitude If the solar radiation scale is converted to optical depth the linear fit to the data indicates precipitation is roughly proportional to e-T 1 3 simulations with the radiative-convective model tested the effect of decreasing scavenging rates thereby allowing the smoke to remain in the atmosphere longer see Figure 8 This has the effect of greatly increasing the maximum temperature decrease and given the stabilization of the smoke l yer that occurs indicates that mid-cont nental temperature changes may be very l irge The fourth simulation tested the effect of removing the cloud cover Such a reduction in cloud cover may result from the warming and subsequent evaporation of the clouds present in the control The effect of removing cloud cover is to further increase the temperature reduction This occurs because without cloudE there will be much less downward emission of infrared radiation from the warm smoky layer toward the cold surface Because the smoke particles are assumed to be small the effect of removing the cloud cover is not made up for by the small amount of infrared emission from the particles The next simulation tested the dependence of the result on surface heat capacity In contrast to the case for land if a high surface heat capacity appropriate for oceanic surface types is used the temperature reduction is much smaller and much more gradual Given that the Northern Hemisphere's surface is about half land and half ocean and that the ocean's large heat capacity 14 Fig 6 Hemispheric distribution of smoke-induced optical depth on day 30 following the hypothetical nuclear er change assuming winte time recipitation cavenging s reduced by e- 3 Hemtspheric average optical depth is 4 5 although these levels are typical only at middle and high latitudes can somewhat buffer the sharp cooling over land we can expect that the hemispheric average temperature change will be somewhere between the land and ocean values probably closer to the latter We may also speculate that the contrasting changes will substantially perturb atmospheric dynamics although this effect will be smaller than expected since the land-ocean temperature contrast will exist relatively near the surface and not at midtropospheric altitudes For the case of no smoke from fires the dust N02 and 03 induce a much smaller perturbation over land than the cases with the smoke and if averaged over ocean and land the effect would be even less The NIJ- l and 03 perturbations only induce a comparatively small change in part because of a cancellation of radiative effects These perturbations are roughly comparable to the effects of major volcanic injections of the past few hundred years and are therefore within the realm of human experience 15 Fig 7 Hemispheric distribution of smoke-induced optical depth on day 60 following the hypothetical nuclear exchange assuming wintertime precipitation scavenging is reduced by e-T 3 Hemispheric average optical depth is 3 0 but these values are exceeded only in middle and high latitudes These perturbations will not be further considered here separately from th e smoke TWO-DIMENSIONAL ESTIMATES 0 CLIMATE CHANGE Instantaneously spreading the injections over the Northern Hemisphere has the effect of exposing meny more of the smoke particles to sunlight than would be the case if this assumption were not made In addition one-dimensional radiative-convective model results clearly indicate the need to treat both land and ocean surface types We have attempted to improve treatment of the effects of spreading of the p nticles and of land-ocean contrast by applying our two-dimensional climate model LSDM In the simulation we have conducted the smoke dust and NQi io jections and 03 reduction are instantaneously spread longitudinally a process that normally takes days to weeks but the 16 40 -- - -- - --- --- 7 JS 30 i j 0 1 0 2 0 3 0 4 Fraction Fig 8 0 5 0 6 0 7 0 8 0 9 1 0 or hemisphere Fractional area of the Northern Hemisphere experiencing optical depths greater than the value indicated at various times following a nuclear exchange assuming wintertime_precipitation scavenging rates are reduced by e-T _ • latitudinal spread that takes weeks tu months is prescribed based on GllANTOUR results The LSDM also as indicated earlier attempts to schematically account for the land-ocean fractionation at each latitude Only very preliminary model results are available from our first two simulations of the hypothetical nuclear exchange In these simulations we have used the GllANTOUR results that assume first normal precipitation scavenging and second that scavenging is damped by a fa tor e-T 3 In this latter simulation our LSDM calculations of the radiative fluxes indicate for example that after a month there is virtually no s lar radiation reaching the lower half of the atmosphere from about 20° N to 70° N - a reduction of about 50-150 W m 1 • Figure 11 summarizes the hemi spheric average land surface temperature changes for the above two cases and the ocean te I perature change for the larger of the two cases For the case with nollllal scavenging we again 11ee that the maximum reduction in land temperature occurs within two weeks ut that it is a much smaller reduction than indic11ted by the o 'e-dimensional mod€1 This is a result both of he modulation of the temperature change by the ocean and of the reduced effect of the smoke 17 sa 1 0 bO C 'i 'iii E Cl e- damping 80 Cl I 0 E Gr e- 3 damping ' E 'O Cl 40 Cl al 2 20 'o Normal sce venging Period of emission 0 0 12 24 36 48 60 72 84 Time days Fig 9 Fraction of smoke aerosol remaining aloft as a function of time for the cases of nonnal and damped two cases wintertime scavenging rates aerosol when latitudinal spread is accounted for It should be noted however that mid-continent al temperature reductions and temperature changes under the smoke cloud which only partly covers the hemisphere could be ubstantially larger Partial recovery of the temperature again occurs over the two to six week interval with the longer lasting change due to the effects of the stratospheric perturbations For the case of damped scavenging the maximwn temperature reduction is only slightly larger than for normal scavenging and much smaller than the -60° C given by the one-dimensional model for the mid-continental land surface type but the reduction persists for much longer than is the case for normal scavenging being still almost as large after three months as is the case for normal sea 1ging after ten days Recovery from the maximum temperature reduction is slowed because there is time for the aerosol to spread more uniformly throughout the hemisphere thereby exposing most of the particles to sunlight and because the ocean temperature is slowly decreasing and therefore cannot as effectively buffer the land temperature _i anges We have examined the effect of the in uced atmospheric stabilization on the precipitation rates predicted by the model for 18 0 -5 -10 0 E- l -15 -20 0 IO 20 30 40 50 60 70 80 90 Time days Fig 10 Land surface air temperature change as a functi ln of the time after the hypothetical nuclear exchange as calculated by one-dimensional radiative-convective model the simulation with damped precipitation scavenging As shown in Figure 12 thare is indeed a reduction in the precipitation rate in this case reaching about 25% over the land and about 20% over the ocean Although this is not as great as an e-T 3 dependence would imply hemispheric average Tat 60 days is about 3 another effect indicated in the model calculations may counterbalance this overestimate Namely the model indicates that the precipitation is more shallow in the perturbed case than in the control simulation thereby reducing the effectivene of the precipitation in scavenging the smoke In addition at the top of the smoke layer the strong absorption of solar radiation induces a warming that would if the process were being represented in our simulation mix the aerosol higher in the atmosphere thereby further reducing the scavenging effectiveness of the precipitation that does occur We have examined the latitudinal pattern of this precipitation change and have found that it is largely a result at one month after the exchange of a reduction in the intensity of the precipitation in the tropical convergence zone This poses further concerns because it is the precipitation in this region that is expected to effectively prevent spread of the smoke to the Southern Hemisphere This issue will deserve close attention in interacti e simulations 19 Table 2 Summary of SensiLivity Tests Using a One-Dimensional Radiative-Convective Model Perturbations Included Soot dust N 03 Sensitivity Tested Maximum Temperature Change °C Time of Maximum Temperature Change days Normal scavenging land surface -30 10-20 Scavenging damped e-T 3 land surface -60 30-60 Scavenging damped e-T land surface -70 50 Scavenging damped e-T 3 land surface no cloud covet -80 30-70 Scavenging damped e-c 3 ocean surface -3 50 Dust NC z 03 only Scavenging damped e-T 3 land surface -7 60 NC z 03 Scaver eing damped e-T 3 lP nd surface -1 60 only SUMMARY Our simulations indicate that significant climatic effects can be expected from the amounts of smoke that have been suggested will be generated from fires started by a major nuclear exchange involving urb n and suburban areas adjacent to or surrounding potential military targets We have not however separately valtlated the estimates of these smoke injections but have instead used the results of other investigators to allow comparison of resulting climatic effects Because the results are so sensitive to emission amount research on this issue deserves attention We have found that there are several important sensitivities related to calculation of the climatic effects Those of highest importance include treatment of the spread and height of injection of the smoke scavenging rates of the soot and of land-ocean differences The effects of changes in cloud cover are also important From th se calculations we can suggest that Northern 20 --- --- --- 0 -1 t-2 § £ Cl Ocean - - Land -3 e-T J damping --- --- ----N orrnal scavenging -4 i - -5 8 -6 E t -7 e-T 3 damping -8 -9 0 15 30 45 60 75 90 Days Fig 11 Reduction in Northern Hemisphere land and ocean temperatures as a function of time following a nuclear exchange assuming either normal or damped e-T 3 wintertime precipitation scavenging rates Hemisphere mid-lacitude land temperatures will cool rapidly days to a week by 10-15° C following injection of the smoke and that temperature changes will be smaller near the ocean and larger in mid-continental regions Precipitation rates will decrease especially in mid-continental regions This precipitation decreas will occur even over the oceans largely because of the atmospheric stabilization of the mid troposphere that shuts off deep convection Absolute temperature reductions can be expected to be larger in su1D1Der than in winter both because there is more solar radiation to be affected in summer and because particle lifetimes are longer in suu mer when mid-latitude precipitation rates are reduced Note however that temperatures in mid-continental regions in summer where the cooling is expected to be largest are substantially warmer than the hemispheric average temperature so that the larger cooling starts from a higher base temperature Circulation and temperature changes will tend to lengthen smoke lifetimes and increase the vertical and latitudinal spread of the smoke 'nlese effects appear to act to further prolong the perturb11tion ACKNOWLEDGMENTS The GRA TOUR model development and applications have been performed in cooperation with John Walton of LLNL W 21 2 14 - ----- 1-- --------------- Normal-Ocean C 0 E - E 12 2 Cl f 8 _ - - - - Normal-Lend Perturbed--Ocean - - - - - - - Perturbed-Land C 0 · al L 4 I Cl c 0 0 30 60 90 Days Fig 12 Changes in Northern hemisphere average precipitation rates over land and ocean for the case of damped e-T 3 wintertime precipitation scavenging rates Lawrence Gates of Oregon State University has graciously allowed use of their GCH output to drive the GRANTOUR model The radiative effects of the smoke have been implemented with the help of Robert Cess of the State University of New York al Stony Brook Others at LLNL including Fred Luthe Jerry Pott r Jim Ellis and Hugh Ellsaesser have helped with recent improvements and applications of the zonal climate model This work was performed under the auspices of the U S Department of Energy by the Lawrence Livermore National Laboratory under contract No W- 740 S-Eng-48 REFERENCES Aleksandrov V 1983 Climatic resp nse to global injections Proceedings of the Third International Seminar on Nuclear War 19-23 August 1983 Erice Sicily Cesa R 1983 The radiative modeling of smoke Proceedings of the International Seminar on Nuclear War 19-23 August 1983 Erice Sicily Crutzen P J and J W Birks 1982 The atmosphere after a nuclear war Twilight at noon Ambia XI 2-3 115-125 Crutzen P J 1983 The illusion of safe areas presented at Third Congress IPPNW Amsterdam June 18 1983 22 Ellsaesser H W 1975 The upward trend in airborne particulates that isn't pp 235-269 in The Changing Global Environment S F Singer ed D Reidel Dordrecht-Holland Gates W L and M E Schlesinger 1977 Numerical simulation of the January and July global climate with a two-leve atm spheric model J Atmos Sci 34 36-76 Hansen J D Johnson A Lacis s Lebedeff P Lee D Rind and G Russell 1981 Climate impact of increasing atmospheric carbon dioxide Science 213 957-966 Jaeger L 1976 Monatskarten des nLederschlags fur die ganze Erde Berichte des Deutschen Wetterdienstes Nr 139 Offenbach Germany 38 pp Knox J B 1983 Global scale deposition of radioactivity from a large scale exchange Lawrence Livermore N2tional Laboratory Report UCRL-89907 Proceedings of the Third International Seminar on Nuclear War 19-23 August 1983 Erice Sicily LamQ H H 1970 Volcanic dust in the atmosphere with a chronology and assessment of its meteorological significance Philosophical Transactions of The Royal Society of London 266 1178 Luther F M 1983 Radiative effects of stratospheric injections Lawrence Livermore National Laboratory Report presented at the Third International Conference on Nuclear War Erice Sicily August 19-23 1983 Manahe S and R T Wetherald 1967 Thermal equilibrium of the atmosphere with a given distribution of relative humidity J Atmos Sci 24 241-259 Penner J E 1983 Tropospheric respon e to a nuclear exchange Lawrence Livermore National Laboratory Report Proceedings of the Third International Seminar on Nuclear War 19-23 August 1983 Erice Sicily Turco R o Toon R Whitten and P Hamill 1982 1-D model simulations of the chemical evolution of the El Chichon eruption cloud EOS Transactions of the Ame1 Geophys Union 63 901 Turco R P O B Toon T Ackerman J B Pollack and C Sagan 1983 Global atmospheric consequences of clear war submitted to Science 23 Walton J J and M C HacCracken 1983 Preliminary report on the global transport model GRANTOUR Lawren e Livermore National Laboratory draft report ' DISCLAIMER Thi documrnc 'as prepar J a an llrrounf of woork on re-d b - an agency of thr llnite-d Sltln Gol-ernmrnf Neither thr United Stat Gowemm 111 nor thr llni er iry of C'alifomi• nor Rny or thrir rmployrri- makes any 111arranlh ex pr or implird or assum6 any leial liability or rrspmnibility for the accuracy complele lfl- or u tefuln·rss of • n information appar alus product or prOCK disclowd or repre i-t'nl'i that hs ould nor infringe prhilh·iy D ned rights Reference he- ein Co any specific commercial products process- or enic e by lradt name trademark m111ufacturer or otbr n¥i le daes not necesuirily constilule or imply it endorsement recommendation or fa•oring by Che Unittd Stair Gotrmmtnl or he Uniwersity of California The iews and opinions of authon e ii pre5 ed herein do not necesurily slate or rened thoJe of the Uniu d S1al6 Gonrnmenl lherttif and shall nol be uH1I for adtertising or product endors t m nt pun_ l5r5 -- __ - - -