JASON - Clouds and Radiation: A Primer (for Climate Control)

8/9/2019 JASON - Clouds and Radiation: A Primer (for Climate Control)

1/29

Clouds and Radiation: A Primier

F. Zachariasen

February 1993

jSR-90-307Op . " for public : ' e : e a s ~ , dlS[l1bullon urllimll . d

JASONThe MITRE Corporation7525 Col shire Drive

Mclea n, Virginia 22102 3481(703) 8836997


2/29

REPORT DOCUMENTATION PAGE Fann "Pf)I"OIi9dOM ' No. 0 1 ~ , ..~ , . . . . , . . . . . . . . . . fortfttt


3/29

Contents1 INTRODUCTION 12 INCIDENT SOLAR RADIATION 73 CLOUD CONDENSATION NUCLEI OVER THE OCEAN 15

II I

I OOe!_S10n rorl ITIS r." r - - - - - - " ~ I . JI hf n:i.C 7 . , ~ ~ i U e g J , , ~ ; , , , ~ n;]; J \ i ' I

t ' ", ) -< r . 1 f'.."


4/29

1 INTRODUCTION

Inclusion of the effects of douds in climate models to a remarkably highdegree of precision is al:nost certain to be a prerequisite to any reasonabledegree of reliability in climate preciiction because of the sensitivity of. forexample, surface temperatures to cloud albedo. Unfortunately, many of themajof r.-hysical processes taking place in clouds are still poorly understood.Some of them involve not only physics, but marine biology, oceanic andatmospheric chemistry, and other disciplines outside of physics. as well. It islikely, therefore, that n o ~ only do physical processes that are understood needto be modeled much more accurately than is now done, out many presentlyignored processes have to be included as well. Indeed, this is the motivationbehind DOE's Atmospheric Radiation Measurement (ARM) program.

This paper addresses, as an illustrative example, one such previouslyunknown complex interdisciplinary process providing a ff'edback loop whichmay have a major impact on the effect on global climate of the steadilyincreasing growth of greenhouse gases in the atmospher('. It is to be stressedthat this is only one of a number of such feedback loops, many of which haveprobably not even been thought of yet, bllt all of which are entirely ignoredil l present day computer models.

Clouds influence climate in two major ways: they reflect incident shortwavplength solar radiation thus preventing it from reaching the earth's surfate an d ::"d.tiull, it. anti they absorb outgoing long wavelength radiation from


5/29

the earth thus reducing the earth's ability to cool its('lf. Which of these twocompeting, opposite sign, effects dominates is a sensitive function of th e in-teraction of clouds with radiation, which is itself a sensitive function of theprocesses going on within various kinds of c l o u ~ s .

Solar radiation is (to a good approximation) a black-body spectrum at.5,780o K, which peaks in the visible at a wavelength of about 0.5 11m. At til('earth, the flux of solar radiation is 1,370 watts/m 2 , and due to the existenceof day and night and because the earth is a sphere, th e average s01ar energyincident at the top of the atmosphere is one fourth of this. Th e average car l Iialbedo is .3, so 30% of the incident radiat.ion is reflected, leaving 240 wattsm 2 to be absorbed by the earth's surface and atmosphere.

Radiation emitted from the earth is (to a relatively poor approximation)a black-body spectrum at 255K, which peaks in the infrared at about 10 11m.There is essentially no overlap between the solar spectrum and the earth'sspectrum. As the outgoing IR radiation passes through the atmospht're. somf'of it is absorbed and reradiated back toward the earth, bu t of collrse th e netIR radiation escaping into space from the top of the atmosphere is also 2tOwatts/m2 , since the earth is (essentially) in equilibrium.

The black-body spectra and the principal atmospheric absorption bandsare shown in Figure I.

The albedo of t h(' ('arth'5 surface (insofar as it can be measured fromsatellites) is no more than .LI). Th e average cloud cover over the earth isobserved to be 60% (65% O V f ~ r thp ocean and 50% over th e land). Hence, to


6/29

(a)

i---01...... ~ I ' m

(b ) 01 10 10011 II I I II I , I I , II

0 , H;O CO, H20 0 , CO, H.,OFigure 1. Curves of black-body energy B, at wavelength h for 5.780 K (approximating to the

sun's temperature) and 255 K (approximating to the atmosphere's meantemperature) The curves have been drawn c f equal areas since integrated over theearth's surface and all angles the solar and terrestrial fluxes are equal.Absorption by atmospheriC gases for a clear vertical column of atmosphere. Thepositions of the absorption bands of the main constituents are marked. (FromR M. Goudy "AtmospheriC Radiation, " Oxford, 1964)

3


7/29

make up the average albedo of .3, the albedo of clouds is about Ar}. Cloudsare therefore the most important component in the amount of reflected sun-light, and their existence is crucial in determ1l1ing th e surface tf'mperatureof t he earth. In fact, (if other parameters were held constant) a change ofcloud albedo by 2% would warm the earth's surface by lOCI. Evidently, ths often hawdiurnal variations which are of considerable importance. For examplf'. th('ratio of stratus to stratocumulus varif's from less than 207c at local noon tonearly 30% in th(' early morning, while cumulonimbus peaks at 30% in lateafternoon or evening and vanishes early in tht' morningl .

4


8/29

Table 1OceanAverage %of Sky Cover over { Land

of Various Cloud TypesCumulus

Cumulonimbus

125

64

Stratus and Stratocumulus :J418

Nimbostratus

Altostratusand Altocumulus

Cirrus

5

66

2221

1323


9/29

2 INCIDENT SOLAR RADIATION

For incident solar radiation, emission is irrelevant because at A " v O..')l1rn.the black body spectrum at the earth's temperature of around 2700K is essentially zero. So only scattering and absorption count. To a first approximation, let us neglect scattering into the beam, and assume th e incidf'ntsolar radiation suffers only losses due to clouds. Thus the radiation intensityI ~ ( k , z) .:>f photons of wavelength>' travelling in the direction k satisfies 1

(2 - 1)

where 0' are absorption and scattering cross sections and . \ is particle density.Therefore,

(2 - 2)where the (dimensionless) optical depth 'T' is

12T{ZZ' z) == O ' ~ ( z ) N ( z ) d z 'I (2 - : ~ ) awl

(2 - 4)is the extinction cross section.

Each type of cloud has , at a given height ZI , a distribution n( r, z) of sizeswith particles of radius r (ice crystals are of course not spherically symmetric.though their absorption cross section does not differ markedly from that ofliquid water droplets). In general, the cross section 0' will also depend on thepartic le size. For wavelengths that are small compared to the particle size,

7


10/29

as typically th e case for solar radiation (see Figure 2) . th e cross sectiollis twice geometrical: (J' = 2r.r2. Thus (if We take .0\ to he th e to p of th eatmosphere where T = 0) th e optical depth at hf'ight :: is

[ top J(::) = 27r iz d::' dr r2 n(r, z'l. (2 - ij)and n(r, ::') is th e droplet size distribution, so that /\'(;;') = .m(r. z')dz'.This equation is usually fe-expressed in terms of th e liquid water content(LWC) of th e


11/29

Table 2LWC (g/m3 ) t (km)

Cumulus .4 2

CumulonimbusTropical 1.0 .JTrade Wind 1.5 2Midlatitude 1..5 2.5Polar 1.5 2

Altostratus/Altocumulus .1 .5

St rat us/Stratocumulus .2

:'-Jimbostratu5 .1 3

average cloud albedo of A,5 results from an average optical thickness r- = 6.assuming a simple geometry of plane parallel clouds. Since in fact clouds arehorizontally variable. and there is some absorption. this value is actuaily alower bound.

Experimentally measured cloud albedos vary greatly, as a function oftype. latitude. solar zenith angle. and other parameter.,. nut overall, theyare not inconsistent with th e inferred value of .4.5. Therefore, we may havesome confidence in the value of T - obtained above.

9


12/29

From tIl(' formula for T ' , we s(( that it varies inversely with til(' dfccti\'('droplet radills. Thus a cloud ha.ving th e same I,W1' as allo\lwr bllt a 1M).';!'\\1l111\wr of smaller droplets, will hav(' a larger T- alld thereforr' a \argPf albedo.

If on(' asks, t h err' fort, , what changes Illay takt' place in dotl,1 aHwdo(to which, as we'\"( rt'markt,d ('arlier. th e earth's sllrfac(' krnperaturt' is px ,tre!lwly scnsitiw') due to man's activities. we must cOllcpnlrate on T' , jJ ' For(xamplt>, an increas(' ill 7" r 11 of 107< (which corresponds to a dt'cn'asc il l ~ . \ " of : ~ O V c ) [edures th e albedo ('!lough so that surfaC(' = 1 . : ~ O ( ' . Or. as illllIther exampk. reducing S by a factor of2 is t'


13/29

25

20 d"""", 135 "md_ " 135 "md'Q = 120 "" '

15


14/29

Manne~ ~ ~ 0o 0 100 20 0:g Droplet Concentration (em"'c:g la)ou'0..,'"".s

Continentall ~ ~ o 200 400Droplet Concentration (em ')

Ie)

'P"I60 0 80 0 1000

Manne40

30

20

100

o 10 20 30Ec:Qg ib )uc:0UOJC. 30 0 Continental25

20 0

o 10 20Droplet Radius (I,m)

Id i

Figure 3. (a) Percentage of marine cumulus clouds with indicated droplet concentrations (b)Droplet size distributions ir. a marine cumulus cloud, (c) Percentage of continentalcumulus clouds with indicated droplet concentrations. (d) Droplet size distributionsin a continental cumulus cloud Note change in ordinate from (b)

11


15/29

clouds, the droplet effective radius, the droplet number, and the number ofCCN, to an extraordinarily high degree of precision; less than 5% accuracywill be required in all of these quantities.

This is an exceedingly stringent requirement on GCMs, and on comput-ing capacity. Nothing like this precision is now available, and, indeed, manyof these parameters are not even included in present models.

13


16/29

3 CLOUD CONDENSATION NUCLEI OVERTHE OCEAN

As greenhouse gases in the atmosphere increase, the earth warms, evap-oration increases, an d the liquid water content of all clouds except cirrusincreases. Therefore, both the albedo and the absorption of outgoing IRradiation increases. Most GCMs predict that the net effect is a positive feed-back, because of the simple observation that an increase of temperature atlow altitude will decrease low clouds while a decrease of temperature at highaltitude will increase high clouds. Various effects, however, may reverse this.

The impact of cirrus clouds is one uncertainty. Their effect is diffi-cult to compute, since they are composed entirely of ice crystals, which areanisotropic and whose effect on radiation is not quantitatively well under-stood. They are also not well studied experimentally; they are very high andalso often even difficult to see visually.

It is also unlikely that fractional cloud cover will remain unchanged ifevaporation increases. Generally, an increase in cloud cover will be a negativefeedback. Figure 47 shows measured changes in average cloud cover over theocean from 1952 and 1980, as a function oflatitucle , annually averaged. Couldthis be due to global warming?

A major uncertainty is the effe(t of global warming on the number ofCCN, particularly oceanic CCN.

l5


17/29

Ocean

+2

... ,

0 4 - + - ~ ~ ~ + - ~ - + ~ ~ + - ~ ~ I ~ O ~ 1 ~ I - + I

-1

-2

60 30 o 30 60 90"Figure 4 _ Change in % cloud-cover from 1952 to 1981 (Annual)

16


18/29

Marine CCN have two major components; sea salt and non-sea saltsulfate (abbreviated NSS). By volume, these are comparable. But the seasalt component is composed of much larger particles, so by number ;-;SScompletely dominates (see Figure 5)8,

To act as ,1 CCN, an aerosol particle must he hydrophilic an d abovea critical size, which is a function of the degree of supersaturation of watervapor in th e cloud. Over the ocean the critical size is thought to be in th erange of .05 to .14 Jtm, corresponding to supersaturation of 0.1 o/c to O ')o/c.(See Figure 5 again.)

Empirically, th e number of CCN over the ocean is around 100 per cm 39(to be compared with tens of thousands per cm3 over polluted land areas).and since this is about the same as th e droplet number density in marineclouds. it is thought that the number of CCN available is a limiting fador inth e growth of marine clouds.

If we accept th e idea that NSS are the dominant CeN, we must next askhow a change in global climate will affect the number of ;\iSS particles. (\Verecall th e apparently very strong correlation with the ice age ending 15.000years ago shown in Figure 6.)9

The present concept is that NSS is produced by the oxidation of variollssulfur ga.'>es coming from th e surface of the sea9 , Dimethyl sulfide ((CH 3 )zS.abbreviated DMS) is alleged to be th e major oceanic source, and it is believed to be of biologic origin. It has a measured con('entration of aboutlOOng/liter at th e sea surface, which varies relatively little over the whole

Ii


19/29

2

1.5

0.5

o 0.01

500

(a) Marine Aerosol Volume DistributiOn

AccumulationMode

0.1Diameter, D (101m)

(b) Marine Aerosol Nurmer DistributiOn

r \ Nuclei mod.'/00 I \I \ I

I \I \

Accumulation mode300

200

100

Diamete" D (Ilm)

6. 0.2 0.1 0.005

Figure 5.

10

0.0002


20/29

CH 3SO; (ng g-')o 5 10 15 -54 -50 -46o ~ - - - " " " T " - - - - - . . . . - - - - . . . , . . . . . - - - -r-.. . , . . . . .--,--,__ .....---. 0

200 5,000

10,000I 400.


21/29

oc{'an surface (by a factor of nf) more than about 2). The entire chain ofmarine biochemistry leading to this concentration of DMS is very complex.involving a number of other molecules and many types of marine animalsand plants. In Figure 79 , displayed is the presently conjedured life c y c J ( ~ ofDMS in the sea.

For our purposes it is not necessary to understand this intricate IH'tworkof connections. The critical question is whether or not oceanic warming wouldincrease or decrease DMS production, and with it, the number of C C ~ andhence both the albedo and cloud lifetime.

Figure 6 suggests a positive feedback; a colder environment produces anincrease in sulfate CCN, which in turn increased cloud albedo which furthercooled the earth. But in fact which way DMS production will change if theearth were to warm is not yet known. The effect could be either a positive ora negative feedback, and, given the extreme sensitivity of the earth's surfacetemperature to cloud albedo, as discussed earlier, the feedback effect couldbe very important.

Evidently, further research is needed on this topic, and very likely itmust be incorporated in some way or other in computer models.

Finally, the existence of this potentially large feedback mechanism. sorecently discovered. suggests that there are other important feedbacks whichare still unnoticed. We are not yet ready for computers to hand us believablepredictions about future climates, even on Vf"ry gross scales.

20

#


22/29

DMSO?

Figure 7.

OMSAN3pM

0.001

?

OMSP-oS-.. ,

21

Phyto.planktonOMSPP

PhytopI/InIdonTot .. 5

SRamineratlzalion

s-.. ,so.28mM


23/29

We would like to thank Stephen Warren and Marcia Baker of the Depart-ment of Atmospheric Sciences at the University of Washington for teachingus the rudimer.ts of the interaction of radiation with clouds, and for providingessentially all of the information summa,.ized in this report.

22


24/29

REFERENCES

1. Slingo, 5., NACAR/0301/89-3, to appear in Nature.2. Warren,S., briefing to JASON (1190).3. See any book on radiation physics, or atmospheric physics, e.g., R. Goody,

1964, "Atmospheric Radiation," Oxford.4. Hegg, X. Q . 198.5, JGR 90 3i33. . Warren. S . 1990, briefing to JASON.6. Tellus 19.58, 10 2.58.7. Anderson. et aL. 1990. as quoted by S. Warren at JASON.S. Charbon, R . et al.. 1987. Nature 326 65.S.

23


25/29

DISTRIBUTION UST

CMDR & Program Executive OfficerU S Army /CSSD-ZAStrategic Defense CommandPO Box 15280Arlington, VA 22215-0150

Mr John M BachkoskyDeputy DDR&EThe PentagonRoom 3E114Washington, IX 20301

Dr Joseph BallCentral Intelligence AgencyWashington, IX 20505

Dr Arthur E BissonDASWD (OASN/RD&A)The PentagonRoom 5C675Washington, IX 20350-1000Dr Albert BrandensteinChief ScientistOffice of Nat'! Drug Control PolicyExecutive Office of the PresidentWashington, DC 20500

Mr. Edward BrownAssistant DirectorDARPA/NMRO3701 North Fairfax DriveArlington, V A 22203

Dr H Lee Buchanan, I I IDirectorDARPA/DSO3701 North Fairfax DriveArlington, VA 22203-1714

25

Dr Curtis G Callan JrPhysics DepartmentPO Box 708Princeton UmversityPrinceton, NJ 08544Dr Kenneth M Case1429 Calle AlturaLa Jolla, CA 92037Dr Ferdinand N Cirillo JrCentral Intelligence AgencyWashington, OC 20505

Brig Gen Stephen P CondonDeputy Assistant SecretaryManagement Policy &Program IntegrationThe Pentagon, Room 4E969Washington, OC 20330-1000Ambassador Henry F CooperDirector/SDIO-DRoom 1E1081The PentagonWashington, OC 20301-7100

DAR P A Library3701 N,Jrth Fairfax DriveArlington, V A 22209-2308

DTI C [2JCameron StationAlexandria, V A 22314

Mr John DarrahSenior Scientist and Technical AdVisorHQAF SPACOM/CN"eterson AFB, CO 80914-5001


26/29

DISTRIBUTION UST

Dr Gary L DenmanDirectorDARPA/DIRO3701 North Fairfax DriveArlington, VA 22203-1714Dr I'\ancy DowdyUSACDA320 21st Street I'\WWashington, OC 20451

Mr John N EntzmingerChief, Advance TechnologyDARPA/ASTO3701 North Fairfax DriveAr Iington, V A 22203-1714Capt Kirk EvansDirector Undersea WarfareSpace & Naval Warfare Sys CmdCode PD-SODepartment of the Navy\Vashington, OC 20363-5100Dr 5 William GouseSr Vice President andGeneral ManagerThe MITRE CorporationMail Stop 26057525 Colshire DriveMcLean, VA 22102Col Randall GressangDARPA/OIRO3701 North Fairfax DriveArlington, VA 22203-17]4

26

Mr. Thomas I i HandelOffice of Naval IntelligenceThe PentagonRoom 50660Washington, OC 20350-2000Maj Gen Donald G HardDirector of Space and SOl ProgramsCode SAF / AQSThe PentagonWashington, OC 20330-1000

Dr Robert G HendersonDirectorJASON Program OfficeThe MITRE Corporation7525 Coishire DriveMailstop 2561McLean, V A 22102Dr Barry HorowitzPresident and Chief Exec OfficerThe MITRE Corporation202 Burlington RoadBedford, MA 01730-1420Dr William E Howard II I 12JDirector of Advanced Concepts &Systems DesignThe Pentagon Room 3E480Washington, IX 20301-0103

Dr Gerald JIafrateU S Army Research OfficePO Box 122114330 South Miami BoulevardRes(>3rch Triangle NC 277092211


27/29

DISTRIBUTION UST

J ASONUbrary [5]The MITRE CorporationMail Stop W0027525 Colshire DriveMcLean, VA 22102Dr George Jordy [25]Director for Program AnalysisUS Department of EnergyER30OERWashington, IX 20585Dr O' Dean P JuddLos Alamos National LabMail Stop A-ll0Los Alamos, NM 87545Dr Bobby R JunkerOffice of Naval ResearchCode 412BOO North Quincy StreetArlington, VA 22217Mr Robert Madden [2]Department of DefenseNational Security AgencyAttn R-9 (Mr. Madden)Ft George G Meade, MD 20755-6000Dr Arthur F Manfredi Jr (10)OSWRCentral Intelligence AgencyWashington, IX 20505

Mr Joe MartinDirectorOUSO(A)/TWP/NW&MRoom 301048The PentagonWashington, IX 20301

27

Mr Ronald MurphyDARPA/ASTO3701 North Fairfax DriveArlington, VA 22203-1714Dr Julian C NallInstitute for Defense Analyses1801 North Beauregard StreetAlexandria, VA 22311Dr Gordon C OehlerCentral Intelligence AgencyWashington, IX 20505Dr Peter G PappasChief ScientistUS Army Strategic Defense CommandPO Box 15280Arlington, VA 22215-0280Dr Ari PatrinosDirectorEnvironmental Sciences DivisionER74/GlNUS Department of EnergyWashington, IX 20585Or Bruce PierceUSO(A)D SRoom 30136The PentagonWashington, IX 20301-3090Mr John Rausch (2)Division Head 06 DepartmentN AVOPINTCEN4301 Sui tland RoadWashington, IX 20390


28/29

DISTRIBUTION UST

Records ResourceThe MITRE CorporationMailstop WI1S7S25 Colshire DriveMcLean, V A 22102Dr Fred E SaalfeldDirectorOffice of Naval Research800 North Quincy StreetArlington, VA 22217-5000Dr John SchusterTechnical Director of Submarineand SSBN Security Program

Department of the Navy OP-02TThe Pentagon Room 40534Washington, DC 20350-2000Dr Barbara SeidersChief of ResearchOffice of Chief Science AdvisorArms Control & Disarmament Agency320 21st Street NWWashington, DC 20451Dr Philip A Selwyn [2}DirectorOffice of Naval TechnologyRoom 907BOO North Quincy StreetArlington, VA 22217-5000SuperintendentCode 1424Attn Documents LibrarianNaval Postgraduate SchoolMonterey, CA 93943

28

Dr George W Ullrich [3]Deputy DirectorDefense Nuclear Agency6801 Telegraph RoadAlexandria, VA 22310Ms Michelle Van CleaveAsst Oir/National Security AffairsOffice/Science and Technology PolicyNew Executive Office Building17th and Pennsylvania AvenueWashington, OC 20506Mr Richard VitaliDirector of Corporate LaboratoryUS Army Laboratory Command2800 Powder Mill RoadAdelphi, MD 20783-1145Dr Edward C WhitmanDep Assistant Secretary of the NavyC3I Electronic Warfare & SpaceDeparbnent of the NavyThe Pentagon 40745Washington, OC 2035Q..5000

Mr Donald J YockeyU /Secretary of Defense For AcquisitionT h ~ Pentagon Room 3E9333Washington, IX 20301-3000Dr Fredrik ZachariasenCalifornia Institute of T ~ h n o l o g y 452-481201 East California StreetPasadena, CA 91125Dr Linda ZallCentral Intelligence AgencyWashington, OC 20505


29/29

DISTRIBUTION UST

Mr Charles A ZraketTrusteeThe MITRE CorporationMail Stop A130202 B u r l i n ~ t o n RoadBedford, MA 01730

29